IET
Decrease font size
Increase font size
Topic Title: High Voltage Breakdown 315KeV
Topic Summary:
Created On: 07 March 2014 01:32 AM
Status: Read Only
Linear : Threading : Single : Branch
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 07 March 2014 01:32 AM
User is offline View Users Profile Print this message



gee706

Posts: 3
Joined: 07 March 2014

Hi,

I'm working with a piece of kit which is used to accelerate ions in the manufacturing of silicon chips. We generate 350KeV using a high frequency power supply and a high voltage stack, basically a Cockcroft Walton Voltage Multiplier.

Bit of advice I'm looking for is based on the fact that the high voltage discharges to the ceiling of the enclosure when raised above 315KeV. The power supply, although rated to 350KeV, is only 0.5mA rated.

I'll try to explain using the photograph and video which I've posted on youtube. The video shows the discharge.

Video of Discharge 315KeV

If the you tube link doesn't work, search you tube for HT400-10DE


The red terminal above is approx 1.4m long, the main box section sits on the red insulators. From red terminal to floor (ground) is 74cm, from red terminal to side walls is 73cm, but red terminal to the ceiling (the ciling is grounded) is 65cm. We see discharge to the ceiling. So, may sound obvious that the ceiling is just too low, but can anyone justify with data to show that 350KeV and 65cm just don't match ? Or do they match ?

In addition to the ceiling gap being so small, I have 1 other thought. I've read about the bloke called Friedrich Paschen, whose Law (curve) proves the relationsip between HV breakdown and pressure. From the top of the red box to the ceiling of the roof, there is an extract pulling 2.9m/s which is required to extract toxic gas fumes.

My thoughts are "to raise the roof" (ceiling really) but I'd be greatful of any technical data that would justify my expense of doing so.
 07 March 2014 11:28 AM
User is offline View Users Profile Print this message



jarathoon

Posts: 1043
Joined: 05 September 2004

I am not an expert on high voltage breakdown, although I do know a little about atmospheric or air conductivity.

http://mycommunity.theiet.org/...unities/files/56/5211

Does it have to be an air gap all the way to the ceiling? What would happen if you affixed PTFE sheet (> 10mm thick) on the ceiling, to cover the exposed metal ceiling? The conductivity of air depends on the concentration of free ions in the air. The conductivity of PTFE (which is very low) remains independent of the number of free ions in the air.

Also is it possible to increase the impedance of your high voltage terminal so there isn't such a big conduction current when the discharge occurs?





-------------------------
James Arathoon
 07 March 2014 10:31 PM
User is offline View Users Profile Print this message



gee706

Posts: 3
Joined: 07 March 2014

Hi James,

I'm initially looking to see if this kit we have had made for us is fit for purpose, as we're not sure if the design was fit for purpose. I can't find other owners of the same kit as its so old, and the manufacturer doesnt support it any more.

I'll take on board the PTFE clading.
 09 March 2014 03:17 PM
User is offline View Users Profile Print this message



dlane

Posts: 690
Joined: 28 September 2007

Generally the dielectric strength of air is taken to be 3kV/mm. There are plenty of texts out there on physics or insulation engineering to support that such as;

Dielectric strength

The dielectric strength will vary dependent upon the humidity and pressure of the air.

Generally if you are sufferring from discharge to surfaces you need to increase the air gap between the 2 conductive surfaces. You can place insulation between the surfaces but this will actually decrease the air gap and will still allow partial discharge to occur when high voltages are utilised.

The alternative is to fill in the air gap completely with insulation material if a larger air gap cannot be obtained. This of course, has its own problems.

Kind regards

Donald Lane
 09 March 2014 05:00 PM
User is offline View Users Profile Print this message



gee706

Posts: 3
Joined: 07 March 2014

Hi Donald,

Thanks for the reply. Here's a photo of the application.

The 3kV/mm would work for me but I think the problem I have is two fold.
The breakdown voltage is also dependant on two things which I don't have:-
(1) Surface condition is perfectly flat, or at least very large radius with no shart points.
(2) Condition of the air - humidity level and also the pressure. Part of the air gap is ducted in a polycarbinate tube for extracting toxic gasses, this then generates a different condition, air is flowing at 3m/s. Electrci breakdown characteristics change as per Paschen's Law.

I have managed to find a University near New York who operate one of these same machines, they have told me two interesting facts about their problems in the past.
(a) Their HV power supply was out of cal, so when they cammanded 350keV, the supply generated excess of 350keV
(b) Whilst not detailing the exact parameters, they had limitations due to poor humidity.

So I've a few things to work on now, I need to educate myself on the best place to be wrt humidity, I need to remove all sharp points, and I need to check HV power supply cal.

Thanks for your help
 09 March 2014 06:24 PM
User is offline View Users Profile Print this message



jarathoon

Posts: 1043
Joined: 05 September 2004

Just some thoughts....

Originally posted by: dlane

Generally the dielectric strength of air is taken to be 3kV/mm. There are plenty of texts out there on physics or insulation engineering to support that such as;



i.e. flat or lightly curved surfaces with no sharp points that unnecessarily amplify the local voltage gradients will help.



The dielectric strength will vary dependent upon the humidity and pressure of the air.


This is the way most electrical engineers are taught to model the electrical properties of air. It is great when you are not interested in modelling what is happening close to or at breakdown initiation - engineers generally want to avoid that.

In regards to standard texbooks - what happens using 50/60 Hz voltages is not necessarily the same as what happens in high frequency breakdown initiation in air.

If you don't have the option of re-engineering the system to avoid breakdown, then you can also consider and model air as analogous to a weak ionic liquid conductor, rather than considering it analogous to a solid insulator. (i.e. a ionic solvent that conducts electricity via the motion of ions. This may be a better way of thinking about things anyway given your ion accelerator application).

Obviously as the discharge itself is initiated this simplistic model becomes invalid, as a cascade of ionisation leads to a hot ion channel plasma being formed. The maximum discharge current will presumably depend on the capacitance and inductance of the chamber, the current supply capability of the voltage source. the impedance of the rest of the circuit and the voltage frequency being used...etc

Ion mobility, in air, decreases as the size and mass of the ion increases. (Fast acceleration of smaller ions may be more important to the breakdown cascades the acceleration of the larger ions, I don't know)

If you add strong alpha or beta emitting sources into the air space the air conductivity will increase, because of the massive increase in the number of positive and negative ion pairs available due to ionisation; you could certainly check the effect of adding lots of small ions with low levels of dust and other pollution in the air this way.

If the discharges in the chamber are initiated like lightning, by up (ion channel) leaders meeting down leaders, then putting PTFE cladding may give an opportunity to reduce the maximum discharge current.

It might be possible to measure the discharge current load on the power supply by adding a secondary earth plane skin to the insulator, and attaching this to the main earth via a calibrated impedance.

-------------------------
James Arathoon
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.