How does it work? Major Revision A.
A few minutes to explain what happens from seconds to femtoseconds.
There are many variations of X-Ray generators from Dental to Baggage scanners to any other form of optical imaging of radiation. The HVDC is ramped up slowly then the filament current is ramped up in the Amp range with DC or RF in a couple seconds then the voltage is ramped up a few more seconds or longer to limit the cable capacitance charge current. The X-rays are formed when sufficicent kinetic energy reduces to time to emit electrons from the anode target in less than a femtosecond \$<0.1*10^{−15}\$ in order to generate the wavelength of X-rays with a frequency of \$>10^{16}Hz\$
XRAY tubes work by using a very high vacuum with some inert gas to raise the breakdown voltage significantly from the Paschen Effect. The tungsten cathode is heated and the anode plate is charged with HVDC pulsed and becomes a big target for the electrons to hit. The target material which is a high temp. conductor instead of a phosphor coating on glass whose X-ray and heat emissions are then re-directed to the next target for imaging. The beam may be magnetically focused like a raster scanned CRT or collimated in parallel to a target or have 2 beams combined for 3D medical imaging.
There is a correlation with filament heater current, cathode to anode current, vacuum pressure and x-ray emission. Often another target called a "getter" collects foreign ion material that contaminates the vacuum and once worn out, there is a deviation in the tube impedance for a given beam current vs filament current or for a given filament current vs beam current.
Unlike previously explained a breakdown Townsend discharge is not desireable as the arc current can be excessive dv/dt, dI/dt inducing stress in the insulation and detonation of foreign materials like carbon deponsits in plastic ethylene polymers or other materials inside the tube.
The cable/connector insulation must withstand this HVDC voltage high stress that may be as much as 5x a colour TV flyback 35kV generator or 150kV as advertised. It must not attract any foreign charged particles or breakdown any gas voids in the plastic as this connector appears to be inserted into the body structure that may be ceramic or glass insulated. The connector may be insulated with oil or deionized water for heat transfer as well.
Since the plastic has a higher dielectric constant than air, the surface must be extremely (class 1) clean and protected to avoid accelerating contaminant gas or particles across the connector insulation otherwise it will create a trail of partial discharges (PD) which can leads to a trail of carbon down the sides of the candle. Meanwhile the tube is charged with a very low capacitance (<<0.1 pF) compared to the cable and connector.
This happens in a picosecond
A Townsend discharge
Due to a high voltage breakdown once ionzation has started when the arc begins, the insulating tube becomes a very low (incremental) negative resistance (R-) collapsing the tube voltage and conducting very high currents, triggering like any gas tube or SCR or ESD pulse except the decay time, dt depends on low capacitance and low negative resistance due to a high I=V/R current thus dt=RC
Some high level x-ray equipment use ramped voltage and filament current and others are steady for continuous imaging.
Only about 1% of the energy is x-ray while the rest of it is heat. So it has a limited life and that life expectancy depends on temperature rise and possibly, daily maintenance to "season" the tube conductor surfaces or burn-off contaminants gently.
So what can cause the plastic insulator candle to build up carbon?
[Speculation]
Partial Discharge of contaminant in the gas, moisture or particles
and high breakdown induced currents in the insulation or a full path Townsend Discharge, which in the question, seems to match the description "Tube Spit"
Insulator dielectrics forces are controlled from E-fields and Conductor magnetic forces are controlled by high current H-fields.
Was it causes by excessive voltage or current for a given contaminant level?
Does higher discharge current translate into greater electrostatic forces on surface contaminants of the cable insulation? Perhaps there were magnetic contaminant atoms around it.
Does the plastic dielectric insulator current rise with HVDC ramp rate?
Yes, so this can affect the reliability of the insulation, but also with the more rapid decrease of voltage by the surface discharge of contaminant gas.
Is the breakdown threshold affect by a dV/dt? YES and NO
- I know from PD test results on insulators that the most PD events occur just before the peak of 60 Hz when the HV is rising, yet from Lightning pulses it takes more time to ionize the zone under stress so it can withstand about twice as much voltage with a microsecond rise time as 60 Hz.
- e.g. a 200BIL insulator bushing ( by definition) can withstand 200kV lightning pulse with 1us rise time and 10 us duration but will arc with << 100kV HVDC or HVAC due to creepage effects on surface contaminants.
Which is less likely to cause ionization or breakdown? DC or impulse?
DC then AC then lightning impulse. Grid components are rated for Basic Insulation Level (BIL) such that BIL200 can withstand 200kV lightning without breakdown yet can easily arc with 100kV AC or DC but is suitable for 40kV grid lines. This is due to surface creepage and E-field gradients for steady-state while lightning impulse E-field gradients have less stress unless it breaks down then much higher H-field current stress.
More speculation ( because they won't give away trade secrets)
I would expect Perkin Elmer ramps HVDC to something like 50kV then ramps the heater current then triggers conduction a 100kV impulse to ensure conduction with a low current trigger spike and high follow-on current stored in polymer capacitance. Then for continuous repetition repeated with cooling by forced air, deionized water or the best with conditioned transformer oil. This way the energy is always controlled by the medium HVDC and the trigger by the HV impulse. This way, continuous 10kW power can be injected with about 1% as x-rays.
Disclaimer: I have no x-ray industry experience per se, but a lot of UHV transformer & RF test experience.
So what can be done to test the components?
Slow HVDC ramp and a PD detector
Test for the PDIV inception Voltage with a slow HVDC ramp until unexpected triggering detected. This is called PDIV. Once you know the ideal BDV threshold, then you know the ideal PDIV, it should be the same. If it is much less then there is an insulation failure somewhere. ( detection is trivial, find the invisible location can be hard) but easier with portable PD detectors with antenna.
Ramp up breakdown voltage with a current limiter and use a scope with a small 1 turn current loop around any insulated conductor via coax and terminated with 50 Ohms. Even a SW or AM radio can pick up the discharge just as it does with lightning 100 miles away. If the current is limited, by series R, then it is a non-destructive test. If any other insulation breaks down such as the plastic candlesticks, this will be detected. But remember, the only thing which causes breakdown to occur less than ideal is contamination or voids.
