Data log of conditioning run
Posted: Sun May 31, 2015 7:34 pm
With all the talk about conditioning lately, I though I'd show some of it's tangible effects. Here is a quick set of data from a run after the chamber is vented. The pressure is measured with a Baratron and cold cathode gauge. The cold cathode gauge does not show up on this large scale linear plot.
The system is controlled and logged through a homebrew SCADA program on a PC.
The chamber is pumped down with a 1402 rough pump while a glow discharge is run. This helps warm things up and sputter some of the crud out of the chamber. The power supply is current limited and the voltage rises over time as the system conditions and the pressure drops. A small 50 l/s turbo pump is then started, extinguishing the glow discharge and putting the supply into voltage limited mode. The system is pumped to about 1E-5 Torr.
Gas is then admitted until a discharge is struck. The pressure is maintained by a mass flow controller on a PID loop with the Baratron and SCADA PC. It is initially set to 7 mTorr with the power supply current limited at 5 mA. An isolation valve before the turbo is manually adjusted to get the system stable in the proper range where the PID can take over. In this run, a relatively low flow rate of about 0.5 SCCM is used which tends to limit the maximum voltage reached. For best performance, opening up the turbo with a high input flow rate will tend to allow for higher voltages but also is a waste of gas if you aren't needing neutrons.
For given a setpoint, the voltage again increases over time as the system conditions. The current limit is stepped up as the system comes into equilibrium.
A few arcing events can be seen as discontinuities on the plot. In general, the pressure needs to be increased to maintain stable operation as the system conditions and the current is ramped up. Increasing pressure also reduces the voltage while in current limited regimes since the effective plasma resistance is reduced.
Neutrons are detectable as the system gets above ~10kV but that is not really the aim of these data. Someday I'll get a ratemeter integrated into the SCADA.
In general, the system will pick up at roughly the same operation points the next time it is run providing it remains under vacuum between runs.
Overview Closeup of beginning of run Even closer up showing some of the initial fluctuations as the system turns on
The system is controlled and logged through a homebrew SCADA program on a PC.
The chamber is pumped down with a 1402 rough pump while a glow discharge is run. This helps warm things up and sputter some of the crud out of the chamber. The power supply is current limited and the voltage rises over time as the system conditions and the pressure drops. A small 50 l/s turbo pump is then started, extinguishing the glow discharge and putting the supply into voltage limited mode. The system is pumped to about 1E-5 Torr.
Gas is then admitted until a discharge is struck. The pressure is maintained by a mass flow controller on a PID loop with the Baratron and SCADA PC. It is initially set to 7 mTorr with the power supply current limited at 5 mA. An isolation valve before the turbo is manually adjusted to get the system stable in the proper range where the PID can take over. In this run, a relatively low flow rate of about 0.5 SCCM is used which tends to limit the maximum voltage reached. For best performance, opening up the turbo with a high input flow rate will tend to allow for higher voltages but also is a waste of gas if you aren't needing neutrons.
For given a setpoint, the voltage again increases over time as the system conditions. The current limit is stepped up as the system comes into equilibrium.
A few arcing events can be seen as discontinuities on the plot. In general, the pressure needs to be increased to maintain stable operation as the system conditions and the current is ramped up. Increasing pressure also reduces the voltage while in current limited regimes since the effective plasma resistance is reduced.
Neutrons are detectable as the system gets above ~10kV but that is not really the aim of these data. Someday I'll get a ratemeter integrated into the SCADA.
In general, the system will pick up at roughly the same operation points the next time it is run providing it remains under vacuum between runs.
Overview Closeup of beginning of run Even closer up showing some of the initial fluctuations as the system turns on