Sciencemadness Discussion Board

Curious glass electrical conduction effect

wg48temp9 - 12-10-2021 at 04:32

I was testing an old power triode with a12kV electronic neon sign transformer to see if the vacuum was good by connecting the 12kV between anode and cathode.

Unfortunately the 12kV was flashing across in the air between my connections so I put a short piece of glass tube of the cathode connection to insulate it. The glass tube was very thin from fluorescent tube with the phosphor removed. The 12kV punctured the glass producing a yellow arc ether side of the puncture. Then on one side of the glass the arc moved away from the puncture leaving two separated arcs on either side of the glass with nothing visible between them as if the glass was conductive. Ater few experiments it was apparent that there was a hot conductive filament between the separated arc points in the glass. But surprisingly the filament was not hot enough to produce a detectable glow.

After the arc was removed and the glass had cooled down the path of the filament was just visible as fine (<1mm) line. Usually the glass cracked or shattered as it cooled down.

Yes I know that molten glass is conductive but I would have assumed it would have to be glowing at least red hot to be sufficiently conductive. Perhaps it was red hot but being so thin the intensity was so low that it was not visible.

I will set up an experiment to take pics of the effect.

Sorry I made a mistake. The electronic neon sign transformer is specified as 10kV, 30mA. That usually means the voltage with no load and the short circuit current.

[Edited on 10/13/2021 by wg48temp9]

Vomaturge - 12-10-2021 at 21:41

I don't think it has a well defined melting point. This video shows conductivity (ion mobility?) before it actually gets hot enough to be a free flowing liquid.

I've laid a glass bottle in a wood fire (not an especially vigorous one) and it slowly deformed into a "puddle," all while acting totally hard and solid when poked with a stick.

Again, it doesn't have to be a great conductor, just enough to pass <100mA with 12 kV across the glass.

Morgan - 13-10-2021 at 12:28

The dielectric strength of soda lime glass is sort of on the cusp of insulating at a mm thickness if your glass was that. And with heat and perhaps some high voltage surface crawling failure was likely. But I don't know how your setup was.

Tsjerk - 13-10-2021 at 12:56

There was a sticky made not too long ago about checking stress in glass. Maybe you have the means to check the glass that way?

wg48temp9 - 13-10-2021 at 21:58

Vomaturge: Yes the glass does not need to be a great conductor.

Given the specification of the transformer is 10kV, 30mA that corresponds to a source resistance of 300 Mohm. Maximum power dissipation will correspond to a load equal to that resistance, Due to positive feedback the conduction increases with an increase in temperature and temperature increases due to heating caused by the conduction. If the feedback is greater than one then the temperature will increase until the the feedback decreases to one. This will result in a temperature such that the resistance is about 300 Mohm, the point of maximum dissipation.

I found a paper on on the electrical properties of soda lime glass at 600C
Attachment: soda-lime-glass-con-.pdf (1.1MB)
This file has been downloaded 101 times

600C is a dark red colour of steel Probably even darker glass.

Morgan: The looks glass I used looked less than a 1mm thick. According to the link you gave, soda-lime glass has an dielectric strength of 13kV/mm The 10kV of my transformer is probably an rms value the peak would be much higher. So 13kV/mm would have been easily exceeded resulting in the puncture happening.

Tsjerk: yes using two sheets of polarizing material. It is possible to peel of the polarizing material on some flat screen monitors. It would be interesting to see the stress develop as the glass cooled.

PS: The electrical conduction of sodium silicate is why at mains voltages heating elements wound on steel tubes and insulated with silicon dioxide and sodium containing ceramic fail at high temperatures.

[Edited on 10/14/2021 by wg48temp9]

Video of the effect

wg48temp9 - 19-10-2021 at 09:12

I dropped a glass mixing bowl which broke. It about 4mm thick with 6mm thick lip. So I thought it would be perfect to try to puncture it with the 10kV.

Initially a brush discharge formed either side of the glass, then after about 15s the an arc formed through the glass and started to create a short filament.

The video below is of an other piece of the bowl. The discharge initially forms a surface discharge round the bottom edge of the glass. The discharge then punctures the glass and forms a filament.

Sorry I have chopped up the video to make it smaller so I can upload some of it. They are only a wnv files.

Discharge round bottom edge.
Attachment: con-glass2.wmv (2.6MB)
This file has been downloaded 112 times

Discharge thru glass and formation of the conducting filament.
Attachment: con-glass-e.wmv (4.7MB)
This file has been downloaded 102 times

[Edited on 10/19/2021 by wg48temp9]

OldNubbins - 19-10-2021 at 10:21

First video the lip of the bowl is intact and the arc goes around. Second video I see two cracks have formed and the arc appears to be going through the one on the right. Then, as the arc on the interior melts the surface of the glass, you can see the arc travel up and leave a path of melted/solidified glass behind it. I don't think the arc punctured the glass, more like the temperature gradient weakened it enough to form a crack which allowed enough ions to flow through and allow the arc to conduct.

wg48temp9 - 19-10-2021 at 13:50


I think you may be right about the cracking. Meaning the glass cracks due to the heat of the surface discharge which then allows the discharge to go thru the glass. This would account for the apparent low dielectric strength of the glass.

Assuming the glass is soda lime glass its dielectric strength should be about 10kV/mm so 4mm thick glass should withstand 40kV. I will be very surprised if a chinese neon sign transformer rated at 10kV is actually more like 40kV.

Though i doubt it now, I originally thought the glass cooking bowl was probably a borosilicate glass which is 20kV/mm to 40kV/mm dielectric strength.

wg48temp9 - 20-10-2021 at 14:40

Apparently it does conduct thru the glass. If its turned off immediately after it conducts thru the glass there is no obvious sign within the glass just a mark on on each surface. Its almost instantaneously thru a pyrex test tube with a wall thickness of about 1mm.

I decided the attempt to measure the voltage using a 10kV 100 megohm resistor but it flashed over. So I tried with a spark gap. It can spark across a 30mm gap which suggests the voltage is at least about 83kV. WOW and its still working. It was probably designed to operate with a neon tube which would limit the voltage.

sparkgap30mm.jpg - 210kB

Junk_Enginerd - 25-10-2021 at 03:06

When it comes to high voltage, glass is significantly conductive >400°C or so to begin with.

But what's most likely going on is due to your high voltage source presumably being AC. AC can be conducted even through infinite resistances due to capacitive effects. Glass has a relatively high permittivity and will because of this gladly conduct alternating currents, more so the higher the frequency is. The glass is acting as a capacitor dielectric, between capacitor plates made up of your plasma electrical arc.

If one puts a sufficiently large capacitor between the poles of an AC source, it is indistinguishable from a hard short circuit.

At a high enough frequency, the whole concept of currents needing to return to ground starts to get weird. A typical example is the air discharges that tesla coils can produce. The electrical current seemingly just goes straight out into the air, despite there being no path to ground. Where to? The answer is that it actually isn't going anywhere except back and forth very quickly. The frequency is high enough that even with a substantial current, it just doesn't move enough charge to deplete even the air before putting it right back again. The air may have a very small capacitance, but at sufficiently high frequency and voltage, it's enough to be significant.

wg48temp9 - 25-10-2021 at 15:40


The initial discharge to the surface of the 4mm thick glass , if it can not reach round the edge is probably due to the capacitance between the two surfaces.

For example a parallel plate capacitor made from soda lime glass 4mm thick with a dielectric constant of 7.7 and an area 1000mm^2 (about 30mmx30mm) has a capacitance of 17pF. At a frequency 100kHz that will have an impedance of about 94kohm, which is small compared to the output impedance of the electronic neon sign transformer assumed to be 330kohm.

I don't know what the switching frequency is but a round 100kHz would be typical for a switching converter. That would be the fundamental frequency, as the output is unlikely to be a pure sine wave it will have lots of higher frequency harmonics.

[Edited on 10/25/2021 by wg48temp9]