Understanding Welding Machines Transients - Doubt

[japasetelagoas]

Hey everyone,
So I work at a locomotive assembly plant and I often hear about locomotives out in the field having their electronics fried up after tasks that involve welding. We do have procedures as to where to ground cable placement, disconnecting all the loco's electronics but once in a while it happens. I do have a few questions that I was wondering if some of you could help me out:

1- Is there a way to measure these transient surges from welding machines? We know that the ground cable needs to be as close as possible to the welding spot, but if that's not possible, how do I assume a safe distance?

2- How powerful can these surges be? Let's say I'm welding some metal bar on one end of the locomotive and I intentionally hooked up the ground cable to the other side of the unit (15 meters apart). Can I estimate a probability that I might fry a bunch of devices along this path?

I apologize if this sounds dumb but it really gets me thinking.

I'd appreciate any good reading suggestions on this topic.

Thanks.

[SeanB]

Just a though, the chances of cooking stuff increase as the loop area increases. Thus, by keening the ground and welding leads close to each other, and then making sure that the loop area is as small as possible means there is a lower curreent induced into the metalwork, and less to flow in direct paths. If you have the leads wandering all over, and around the loco for the one lead and the ground the other side you will cook something, but running them together will reduce this.

[CatalinaWOW]

I like the loop answer, particularly if your welder is one that has a high frequency arc initiator feature.  There is another factor which could contribute.  Welding currents are enormous.  200 or more amps.  A mere 0.1 ohm of contact resistance theoretically could give 20 volts.  When you put your ground far away there is no telling where that current will go as it seeks the path of least resistance (pun intended).  Applied to the wrong places 10 to 20 volts could easily fry electronics.

[T3sl4co1l]

Besides impressive DC currents, the transient voltages can be quite large.

HF start may not deliver much current (a fraction of an ampere, RMS), but if the spark is make-and-break rather than a continuous arc, each individual spark event causes a discharge of the welding cable's capacitance into the surrounding material.  It doesn't sound like much, but the cable contains 10s of nanoseconds of charge, and exhibits a high-frequency impedance on the order of 100 ohms.  Starting from a high voltage like 10kV, that's I = V/R = 100A!  If even a fraction of this energy gets into an electrical signal pin (like a communication port or logic device), it can damage or completely blow out the minuscule structures inside the chips.

During a steady arc, the transient voltage should be fairly small, or no worse than the high frequency energy delivered by maintained HF start, or AC (inverter) drive.  The instant the arc is broken, however, the voltage jumps up suddenly, across all conductors in the loop (proportional to their inductances).  And since the loop was carrying full welding current (whatever that happened to be -- which might not be very much for a TIG/GTAW process,  where the current is tapered off at the finish of a pass, down to 10s of amps), the voltage thus generated has a low impedance, so that the current capacity behind the transient is comparable to the welding current itself.  So, a peak of ~100s of volts and up to 100-200A, let's say.

It seems weird to me that something as big as a rail car, and needing to be as reliable as it does, would be designed without electrical systems that can handle this kind of abuse.  Signal isolators and transient protectors do add some cost, but when the production quantity is small and the reliability must be high, one shouldn't miss the opportunity to add such things!

Tim

[japasetelagoas]

Besides impressive DC currents, the transient voltages can be quite large.

HF start may not deliver much current (a fraction of an ampere, RMS), but if the spark is make-and-break rather than a continuous arc, each individual spark event causes a discharge of the welding cable's capacitance into the surrounding material.  It doesn't sound like much, but the cable contains 10s of nanoseconds of charge, and exhibits a high-frequency impedance on the order of 100 ohms.  Starting from a high voltage like 10kV, that's I = V/R = 100A!  If even a fraction of this energy gets into an electrical signal pin (like a communication port or logic device), it can damage or completely blow out the minuscule structures inside the chips.

During a steady arc, the transient voltage should be fairly small, or no worse than the high frequency energy delivered by maintained HF start, or AC (inverter) drive.  The instant the arc is broken, however, the voltage jumps up suddenly, across all conductors in the loop (proportional to their inductances).  And since the loop was carrying full welding current (whatever that happened to be -- which might not be very much for a TIG/GTAW process,  where the current is tapered off at the finish of a pass, down to 10s of amps), the voltage thus generated has a low impedance, so that the current capacity behind the transient is comparable to the welding current itself.  So, a peak of ~100s of volts and up to 100-200A, let's say.

It seems weird to me that something as big as a rail car, and needing to be as reliable as it does, would be designed without electrical systems that can handle this kind of abuse.  Signal isolators and transient protectors do add some cost, but when the production quantity is small and the reliability must be high, one shouldn't miss the opportunity to add such things!

Tim

Can't thank you enough for the reply!

[eKretz]

I guess my first question would be why wouldn't it be possible to place the ground right next to the weld zone? It is always best practice to place it as close as possible, and even to disconnect power wires to sensitive components/batteries if possible. If it's simply a matter of access to a clamping point for the ground - there are many options for tough areas - including magnetic grounds.

I have even made up a ground lug from copper with a rare earth magnet pressed in to use for these situations myself. It's around a 1¼" diameter copper cylinder bored out and with a 1" neodymium magnet pressed in the back side. The flat copper side goes against the work, and there's a lug welded to the copper on the back for a ground clamp to bite on. Works great.