The IceCube PMT Anode Circuit Transformer
The transformer consists of a number of turns of thin coax cable on a
high μ ferrite core. The primary of the transformer is the center
conductor of the coax, the secondary is the coax shield.
The transformer solution serves IceCube in several ways:
The short length of coax stores little electrostatic
energy while giving wide bandwidth. A capacitor coupled output
with comparable bandwidth would store about 1000 times more energy.
The transformer yields some common-mode isolation in
the DOM MB input circuit.
The image below was produced with an IceCube (and AMANDA string 18)
style transformer driving a 95.3 ohm resistor load monitored by an
active oscilloscope probe.
The input circuit waveform is observed at a "T" connector several ns
from the transformer. The impedance of the cable itself is
evident
from the rising edge to 3.2 divisions across the screen.
Then, the response of the transformer is evident. Historic Pix1, Pix2
from prototype tests durin 1999... Some droop is evident.
The spike in the input waveform (red waveform) at 3.2 divisions is
inductive reactance (possibly leakage inductance) of the transformer.
Since the input impedance of the transformer is higher than the
impedance of the coax, a parallel resistor is used to 'match' the
transformer to the line.
A flaw in the set-up results in a turns ratio that is not unity, so the
parallel resistor turns out to be a value that is higher than expected
by naive analysis (143Ω)
In spite of all its warts, it does have a rather fast rise time, and
seems to be relatively free of shock-induced oscillations (ringing).
For now, consider the above to be preliminary results which are,
nevertheless, encouraging.
Future plans for tests:
When time permits, the impedance test will be redone with a transformer
configuration that better matches the expected implementation.
- the center conductor will be driven by the input pulse instead of
the shield (i.e. this test was done in the reverse direction)
- the cable impedance will be more accurately measured so
impedances can be more accurately calculated.
- active probes will be used at the input and output of the
transformer to confirm the waveforms
Input and output impedances will be adjusted until the 'best' waveform
is coupled through the transformer. This will allow us to
determine whether we have made the right implementation choices, and
whether the satisfactory operating range/latitude is wide.
To be added to this page:
Photos of completed transformers.
Bill of materials for the transformer.
Gerald Przybylski
March 12, 2003
Lawrence Berkeley National Laboratory