Understanding Plasma THC

[nerginer]

Hi,
I want to understand the logic behind the Plasma THC system. I have a few questions.

1- Is there a arc ok output connection in powermax1000-1650 systems or do I have to use a transformer
2- What is the output voltage of this arc ok output
3- Is there any schematic in order for scale and filter tip voltage to 0-5v range
4- Is there anybody who want to develop his own solution for THC

Regards,
Nuri Erginer

[samakthachai]

Hi Nuri,

Go to http://groups.yahoo.com/group/CandCNCSupport/.

You may clear for your problem.

Cheers,
Samak

[Robin Hewitt]

Originally Posted by nerginer 4- Is there anybody who want to develop his own solution for THC

I've been looking at it but my plasma table got stuck on hold while I build something else. Work takes precedence over fun I guess

My cutter fires strike pulses the moment the arc is not okay so they would have to be filtered out, I reckoned it was easier to pick up on them than wire in to the box.

Two filters required, high pass for the strikes and low pass for the torch height voltage.

Dropping the volts to digital levels would be a pair of resistors, but do you want to drop it all the way? You are trying to detect a fairly small voltage shift so perhaps you want to bring the ADC ground up to the arc rather than scale the whole lot down and lose resolution.

I really need to get my 'scope on this to get a feel for the numbers before I can progress. Love to see some 'scope traces if anyone has them

[nerginer]

>My cutter fires strike pulses the moment the arc is not okay so they would have to be filtered out, I reckoned it was easier to pick up on them than wire in to the box.

We will wait for the arc ok signal from the plasma before starting the THC. This may eliminate some strikes but offcourse we have to use filters.

>I really need to get my 'scope on this to get a feel for the numbers before I can progress. Love to see some 'scope traces if anyone has them

You are right! We have to collect some data in order to design the filters. Is there anybody who can do that?

>Dropping the volts to digital levels would be a pair of resistors, but do you want to drop it all the way? You are trying to detect a fairly small voltage shift so perhaps you want to bring the ADC ground up to the arc rather than scale the whole lot down and lose resolution.

We can use pic or atmel based cheap microcontroller's ADC which works between 0-5v with high resolution so I hope we can not lose resolution

Lets talk more about the subject later we can use a blog to share the files.

[Robin Hewitt]

Originally Posted by nerginer We can use pic or atmel based cheap microcontroller's ADC which works between 0-5v with high resolution so I hope we can not lose resolution.

My point was that if you take a 100 volt arc (I'm just pulling numbers out of the air here) and divide by 20 to get 0-5 volts you could end up looking for changes between 4.2v and 4.3v.

Theoretically possible if you can connect to both ends of the ADC resistor ladder, but would there be a problem screening it from the huge voltage inverter rattling away right next door

[nerginer]

please take a look at these sentances taken from the rutex user manual

The arc voltage must be externally divided into less than 5 volts. This signal is applied to ARC_VOLTAGE_1
and _2. The difference between arc voltage and reference voltage is amplified and fed to the servo driver.
The speed of the movement is proportional to the difference of these voltages. The reference voltage is a
preset value of the arc voltage, 0 to +5V, delivered from panel mount potentiometer or analogue output from
the CNC.

Selection of voltage divider for plasma arc.
The function of voltage divider is to divide the arc voltage in ratio that the output voltage from the divider
does not exceed 5V. When THC operates correctly and certain voltage (REF J2 pin 4) is preset, the height of
the torch is adjusted so that reference voltage (J2 pin 4) is equal to divided voltage from arc.
For example a 50:1 divider is used and required arc voltage is 125 volts. The reference voltage must be
preset to 125 / 50 = 2.050 volts.
When using 953-002 PCB, which has 5V reference diode on board, with standard 10-turn multi turn
potentiometer with 15 turns dial a voltage can be directly read from dial of this pot. The low side of this pot
must be adjusted to 1 volt, high side to 3 volts and dial set from 50 to 150. With 50:1 divider, the preset
voltage 2.050 will be read on dial as 125 +/- tolerances.
When selecting values of resistors for voltage divider an input impedance of 20-kilo ohms on ARC
VOLTAGE inputs (across pin 12 and 14 of J1) should be taken in account in case of use high value resistors.
The output of the divider should be bypassed by the capacitor (in ten or hundreds of nanofarads) to
attenuate high frequency from arc and protected by a high-speed low voltage transient suppresser (such
General Instrument SA5.0, SGS-Thomson BZW03P5V8, AVX-Kyocera VA100005A150).
A standard 50 : 1 divider is also available from Rutex (p/n 983-000).

[Robin Hewitt]

So you preset the voltage you want on the arc and it moves up and down to try and maintain it? Not exactly trying very hard are they

I think I prefer set the torch height, start moving, measure the volts before things start to bend and then stick with that.

However self calibration to height does mean knowing when the tip touches down and knowing that it hasn't bent thin metal underneath it

[nerginer]

the flow chart can be like this

[Mike@Torchmate]

The good news is all these answers have been solved because there are a few different AVHC units manufactured. The bad news is, no one will tell them to you.

[Torchhead]

You have to be able to read and react quickly to +- 1 V change at the tip. So in a 50 to 200VDC range you need to pick out a 1 volt change with the tip moving 200 IPM and 20Volts of noise. More than 3 volts out and your cut starts to sputter or the tip bangs the metal. Your feedback and adjustment become a servo loop. Filter too much and you get poor response. Filter too little and it becomes a random position control. If you don't understand grounding and isolation you will burn up a lot of circuits.

If you can't afford the 350.00 THC300 I doubt you have the equipment you need to build and test a working ATHC.

It's simple until you get to the details. I spent MANY hours of engineering time to get a working THC and then many times that amount in testing and tweeking for something that is reliable and will survive the plasma cutting environment.

We are close to releasing our third generation of THC based on hundreds of installs, countless hours of testing, and even more doing customer support.

Funny you should quote the Rutex documentation. They never could make that work good enough for prime time use. It was never more that a "kit" and their approach to pierce height sensing was a joke.

Tom Caudle
www.CandCNC.com

[Robin Hewitt]

Hi Tom

I have no desire to make commercial THC electronics or publish DIY THC plans. I have all the equipment but it's a minefield of patents and I do occasionally like to visit the US of A without going to court and having my buns sued off

OTOH I want to add THC to my table, should I ever finish it, but the design is a bit unorthodox and I think the THC300 rather assumes I would be doing it the same as everyone else.

If I can get simple up down instructions out of a box which I can relay to the cutting head down a busy fibre optic link, then 350 bucks would be money well spent.

Respect

Robin