All in the PSU
A lot of early development went into the power supply design. I was
particularly proud of this custom wound 3kV transformer which provided me with service through much abuse and
many further development stages. It is around a 6 inch cube and good for up to 300W continuous.
Attached to a number of different cockroft-walton multipliers it provided me with supplies up to
15kV. I tried to be as gentle and safe as possible with this by using a Variac on its input from the mains to
allow me to gently raise the supply on each power up.
From the small table top arrangement that the previous arrangement was, the supplies based on this transformer
were all very much floor based as the multipliers got larger and larger.
To start with I smoothed the original 5kV supply with polypropylene
lighting caps. I found that as I started to build larger arrays of plasma electrodes the current use
increased the ripple to audible levels and the supply voltage sagged. This was an early attempt to get
around it. What tends to happen when a plasma is struck is that once at the strike voltage the current
consumption then loads the supply and if it isn't man enough the voltage dips and the plasma switches off
again in a repeating cycle.
Then with the new transformer I utilised some high voltage
electrolytics for the multiplier with the same smoothing. These electrolytics were disastrous - although
specced for the voltages with balancing resistors they couldn't handle the current stress and exploded
regularly, this does nothing good for the nerves. The whole multiplier is housed in a polycarbonate
shell here with transparent lid - purely for shrapnel protection (it's about 1m wide by 1.5m long).
So having learnt I stuck with polypropylene
from now on and built a complete 5 stage multiplier with the lighting caps. Each blue cap is 15uF, 630V
- they are arranged in series sets of 5 giving me an equivalent of 3kV and 3uF each. Shown here
with a stereo pair of 10kV amps I'll come to later. This was a really stable and strong arrangement,
polypropylene caps can survive anything - a bit more than the diodes anyway. I lost count of the number
of times I replaced the string of 1 amp diodes in the centre of this. Whenever an accident happened on
the output it's those that took the brunt of it.
I used this supply for most of the following experiments. The 3kV AC was rectified and multiplied to
around 15kV DC. However, depending on the loading I suspect this voltage varied considerably. Besides
that disadvantage this type of supply has a number of advantages. With no load the voltage was more like
18kV, useful for striking arcs. With that size of capacitors it could withstand the start-up surge of the
plasma striking with no problem. There was very little other noise apart form 100Hz ripple - but as the
plasmas were no good at bass that didn't matter for a number of experiments. It provided up to 300W
continuous power and really didn't mind being overloaded so long as it wasn't actually shorted. So for
experiments it did me very well. Of course it was also lethal and I always treated it with utmost respect,
often powering up with the Variac from 3 metres away. I get the feeling people thought my careful checks and
nerves before each switch on were for show.