6. AC TUNING
Please read section 4, "General Guidance on Pulse-Shaping", before trying to optimize the pulse-
shape of your unit. Note that all pulse tuning should be performed with the actual output cable you
will use.
6.1
The top of the output pulse should be quite flat. However, when using an RF sweeper at a
fixed frequency, we have often seen a tilt in the output pulse over the first few microseconds. This
is an artifact of the sweeper, and the phenomenon is highly frequency- dependent. Therefore, if
you observe a tilt over the first few microseconds, and are using a sweeper (set at a fixed
frequency) to generate the RF, adjust the sweeper frequency by a small amount to see if the tilt
changes. If the tilt does change, adjust to a frequency which minimizes the tilt (or put in an
isolator). If a significant tilt persists, it is prudent to examine the detector output pulse directly, to
see if the tilt is caused by the RF (e.g., by poor RF grounding). This is done by turning off the
amplifier power supply, and examining the waveform at the input of A1. In our experience with
many L-17C units, any tilts, undershoots, etc., which might be associated with the IC itself must be
very small (much less than 1 dB).
6.2
Turn the RF power to less than -60 dBm to make sure there is no feed-through from the
modulator switching circuit. This can be minimized by using as small an input pulse to the
modulator as possible, and by placing attenuators between the modulator and the detector. Poor
pulse-shape at low power is often caused by poor modulators.
6.3
We have observed a small overshoot on the output with a time constant of ~100
nanoseconds. This can be removed with an RC net (R18, C17) [Fig.2(c)], provided the overshoot
is linear, i.e. the same percentage overshoot occurs at all powers. Typically, the values R = ~45k
and C = ~2 pF will remove this overshoot.
6.4
If you have an overshoot at low power only, this can be corrected with an RC net on A2 or
A3, as appropriate.
6.5
If you have a long, lossy cable on the output, and the output pulse initially rises rapidly and
then has a slow rise to full power, proceed as described in section 4.2, "Good Pulse-Shape With
Lossy Cables".
6.6
The rise time is surprisingly constant over power. Lower powers do, however, tend to have
longer pulse transit-times. If it is important to keep the rise time constant, or to minimize transit-
time variation, low- and intermediate-power signals can be selectively adjusted by speeding up, or
slowing down, A2 and A3 as described in section 2.3. There is little that can be done with A1.
6.7
If the trailing edge shows any ringing or hooks, the problem is probably ringing in A1. (In
general, this problem is due either to an incorrect value of C2, or poor front-end layout.) Try
increasing C2 or C4. In extreme cases, you might replace C4 with an RC combination with R = 80
ohms.
-18-
 |
 |
 |
 |
 |