Blowing up coax cables

[dpal]

Hi! I am powering an inductively coupled plasma with a 13.56MHz RF PSU at around 300-500W. It is matched with an L matching network such that <10W are reflected, and connected to a water cooled 4-turn copper coil. It runs stably for maybe 20 minutes, but then melts through the several mm of PTFE in the coax between the coil antenna and the matching network (MN side) and shorts.

https://imgur.com/a/o83wRFK

(this is just the insulation, ignore the red wire - it's the same with standard coax connectors)

This happens even with 0 reflected power and with thick insulation in the connectors. Sometimes it just happens after a certain time, sometimes as I'm slowly changing gas composition from argon to hydrogen.

What should I look into to diagnose the root cause?

 

[mtwieg]

What you're showing isn't a coaxial cable/connector, or at least not one I recognize. Need more information.

Is the matching network on the load side of the cable?

 

[vfone]

PTFE has high melting temperature of 327°C (621°F). Even when molten, PTFE does not flow due to its exceedingly high melt-viscosity.
Maybe the coax dielectric is not PTFE (Teflon).

 

[volker@muehlhaus]

in the coax between the coil antenna and the matching network (MN side)

So this is NOT in the 50 Ohm path?

Any idea what the input impedance into the coil is? I guess you might have really low impedance = large currents there, and due to skin effect, all that high RF current pushes towards the surface of your wires and connectors.

 

[D.A.(Tony)Stewart]

Plasma arcs have negative resistance which means with a low capacitance electrode the spectrum is very high in the microwave band and may be stochastic which should disrupt all nearby mobile communication as well as create surface high power density from skin effects. If the connector gold plating is removed from improperly torques or insufficient with oxides the conductor surface will heat up. Although you are mixing inert gases, and your coil return loss is excellent, the plasma arc currents will disrupt the crest factor of losses to not be sinusoidal.

500W into 50 Ohms implies almost 200Vp under linear conditions is not the problem, rather the kV from flyback energy.

I suggest you upgrade your connectors as one would for broadcast power levels. (e.g., RG-393, SFT-500) with Type N, pref. HN, or 7/16 DIN connector with silver plating for low loss.

Poor thermal management and the high voltage from flyback will cause plasma generators to burn. I recommend air flow and semi-rigid copper coax.

You can listen to the plasma modulation using a tuned-circuit detector to learn how to control your combustion noise levels with temperature, gap control, inert gas medium and power levels.

This is useful for C60 production too but 4% is the LEL for H2 detonation will be a critical tuning threshold.

 

[dpal]

It’s a deconstructed female coax output from the matching box. Here’s another picture:

https://imgur.com/a/3sUP01X

Hopefully that also attaches the schematic of the system

--- Updated Apr 6, 2025 ---

It’s a deconstructed female coax output from the matching box. Here’s another picture:

https://imgur.com/a/3sUP01X

Here’s also a schematic of the system

Plasma arcs have negative resistance which means with a low capacitance electrode the spectrum is very high in the microwave band and may be stochastic which should disrupt all nearby mobile communication as well as create surface high power density from skin effects. If the connector gold plating is removed from improperly torques or insufficient with oxides the conductor surface will heat up. Although you are mixing inert gases, and your coil return loss is excellent, the plasma arc currents will disrupt the crest factor of losses to not be sinusoidal.

500W into 50 Ohms implies almost 200Vp under linear conditions is not the problem, rather the kV from flyback energy.

I suggest you upgrade your connectors as one would for broadcast power levels. (e.g., RG-393, SFT-500) with Type N, pref. HN, or 7/16 DIN connector with silver plating for low loss.

Poor thermal management and the high voltage from flyback will cause plasma generators to burn. I recommend air flow and semi-rigid copper coax.

You can listen to the plasma modulation using a tuned-circuit detector to learn how to control your combustion noise levels with temperature, gap control, inert gas medium and power levels.

This is useful for C60 production too but 4% is the LEL for H2 detonation will be a critical tuning threshold.

Thank you! The disruptions are mitigated with a Faraday shield made of a copper mesh. I’ll try adding a flyback diode and getting a copper cable (looking at this one https://www.fairviewmicrowave.com/p...gid-coax-cable-copper-fm-sr250cu-st-bulk.html). We are already cooling that cable with 3 fans. I’ll check what connectors we have in the lab, but pretty sure I used something like RG-393.
I’m not sure what you meant in the last paragraph, we don’t produce C60, and the plasma is up to 100% hydrogen at 0.1torr.

 

[D.A.(Tony)Stewart]

Time to upgrade connectors and coax with larger types.

if on a tight budget

PL259 UHF

or bigger budget

or

 

[FvM]

It would be helpful to have rough estimation or measurement of load impedance. Also state of matching box (estimate value of L and C) components can show it.

For the time being I agree with Volker that it's more likely a problem of cable and connector current rating.

 

[biff44]

that is not "melting thru" the cable.
that is arcing. an arc forms, then all bets are off as the plasma errodes the teflon.