Hypertherm PMX ARCOK signal latency

Hypertherm Plasma Cutter discussion forum.
Post Reply
robertspark
4.5 Star Elite Contributing Member
4.5 Star Elite Contributing Member
Posts: 1806
Joined: Mon Jun 12, 2017 6:43 pm

Hypertherm PMX ARCOK signal latency

Post by robertspark »

Hello,

I was wondering if Jim Colt or anyone else who may know could help.

What is the ARCOK signal latency / delay from the time that the arc is actually lost (resume there is some debounce here to confirm that the arc actually is lost and it's not a false signal), to the time that it is output from the Hypertherm PMX range of plasma cutters.

Note I have a PMX 45 (non-XP) and would be most interested in my specific case, but am also more generally interested roughly what we are looking at.

I am aware that the PMX plasma cutters output the ARCOK signal via dry contact (i.e. relay closing) and presume that you've selected some fast operating signal relays mSec closing and opening range).

Most of the fast acting relays I've seen have closing times in the 3, 5, 10mSec range and opening times in the 1-2mSec range so arc lost would be really fast, and arc OK (relay closing) a little slower.

Any guidance would be appreciated thanks from Jim / hypertherm technical or anyone else in the know wish some numbers.

We have a discussion on another forum about latency of motion controllers and external THC's receiving ARCOK signals in the event of the arc being lost and the motion controller (or external THC) being able to recover to the exact position of where the arc was lost, considering the feedrate at the time the ARCOK changed state, and any latency / debounce not only within the plasma cutter (machine side) but within the motion controller and external THC (which is easier to determine because I am talking with both the motion controller and the THC developer)

Thanks in advance

Rob
tcaudle
4 Star Elite Contributing Member
4 Star Elite Contributing Member
Posts: 1353
Joined: Tue Feb 28, 2012 6:47 pm

Re: Hypertherm PMX ARCOK signal latency

Post by tcaudle »

Why don't you measure the actual arc current if You want to be accurate? Most arc ok circuits have to have some delay built in to avoid false dropouts Typically relays take longer to drop out than to turn on because the coil voltage keeps it on down to a lower voltage than it takes to turn it on (holding voltage) so a 12V relay may pull in at 10V and dropout at 6, Relays are not very good real time signal devices where speed or consistency is needed.
Each plasma cutter will vary . Some may not even have conventional coil relays for Arc Transfer/ Arc oK/ Ok to Move.

It sounds like you are trying to do an auto restart of a cut. Interesting but its got a lot of challenges. Going back to the exact position of where the arc ok signal drops out may or may not be where you want to try and reestablish a cut.

One handy option is to be able to click on a cut segment in the toolpath display and it jumps to the G-code line (and vice versa) Sometimes you can replace an operator with software.
jimcolt
5 Star Elite Contributing Member
5 Star Elite Contributing Member
Posts: 3087
Joined: Mon Jul 20, 2009 11:18 pm
Location: North Carolina

Re: Hypertherm PMX ARCOK signal latency

Post by jimcolt »

Hypertherm has always (well for the last 49 years) called this signal the "Arc Transfer" or "Machine Motion" output. We use a small, fast, mechanical relay as there are a few hundred different cnc machines worldwide that this signal interfaces with.

When a start signal is issued to the plasma cutter from the cnc, and if the torch is in close proximity to the material to be cut...the torch pilot arc fires, then the arc finds a better path (from the negative arc to the positive side of the plasma power supply...through the work lead attached to the material) to the material. As the power supply ramps up (different processes have different up ramp curves) the current increases on the work cable and when cutting power is sensed (current sensing) on the work lead we close the arc transfer relay.

The arc transfer relay contact is monitored by the cnc to determine if the cutting arc is actively cutting metal. This is used to start a pierce delay timer that allows adequate time for the plasma arc to fully pierce...then the THC should first index from pierce height to cut height...then the CNC should start x and y motion.

The arc transfer relay is also monitored while cutting in case the arc extinguishes for any reason, so the cnc can stop motion, height control, etc.

If you need technical details on how fast the relay manufacturer claims their relay will open or close.....or how fast we expect our current sensing to occur (which likely would be different for each different Hypertherm model for a variety of reasons!)....you could contact technical service at Hypertherm. I'm not sure if they can provide you with timing related answers, but they are the best source. technical.service@hypertherm.com .

Jim Colt


robertspark wrote:Hello,

I was wondering if Jim Colt or anyone else who may know could help.

What is the ARCOK signal latency / delay from the time that the arc is actually lost (resume there is some debounce here to confirm that the arc actually is lost and it's not a false signal), to the time that it is output from the Hypertherm PMX range of plasma cutters.

Note I have a PMX 45 (non-XP) and would be most interested in my specific case, but am also more generally interested roughly what we are looking at.

I am aware that the PMX plasma cutters output the ARCOK signal via dry contact (i.e. relay closing) and presume that you've selected some fast operating signal relays mSec closing and opening range).

Most of the fast acting relays I've seen have closing times in the 3, 5, 10mSec range and opening times in the 1-2mSec range so arc lost would be really fast, and arc OK (relay closing) a little slower.

Any guidance would be appreciated thanks from Jim / hypertherm technical or anyone else in the know wish some numbers.

We have a discussion on another forum about latency of motion controllers and external THC's receiving ARCOK signals in the event of the arc being lost and the motion controller (or external THC) being able to recover to the exact position of where the arc was lost, considering the feedrate at the time the ARCOK changed state, and any latency / debounce not only within the plasma cutter (machine side) but within the motion controller and external THC (which is easier to determine because I am talking with both the motion controller and the THC developer)

Thanks in advance

Rob
robertspark
4.5 Star Elite Contributing Member
4.5 Star Elite Contributing Member
Posts: 1806
Joined: Mon Jun 12, 2017 6:43 pm

Re: Hypertherm PMX ARCOK signal latency

Post by robertspark »

Thanks for you both taking the time to post.

Yes, the discussion is regarding automatic cut recovery. It does work with uccnc and with the neuron thc, the discussion was regarding latency in the system between machine sense and output, THC input and output and motion controller input (which with the uc motion controllers is very little).. plus the effect of breakout board s on signal latency. With a view to improving the latency of arc transfer signal forwarding between the neuron at uccnc (which I think we now have a way and it is being looked at by the neuron developer Andrew) and also allowing for arc transfer machine side signal latency which we may also have a way to improve too

Thanks again for your input

Rob
Post Reply

Return to “Hypertherm Plasma Cutters”