-3
resistance. The number thus obtained is the criterion for the selection of the slope-matched
pairs.
Of course, since the absolute value of the slope-matching parameter will depend on the detector used, exact
power levels, etc., it will be best to select pairs using only contemporaneous measurement data.
C: Circuit Error in the L-17C Application Notes
We just recently discovered that a last-minute change in the L-17C circuit has been incorrectly documented
in all revisions of the L-17C application notes. Pin 10, which is designated "L6 ADJUST" in Fig. 1(b), should
be labeled "L7 ADJUST". Thus R110, referred to in Fig. 2(c) as "LOG 6 CURRENT ADJUST", actually is
"LOG 7 CURRENT ADJUST". Please make the appropriate changes in your application notes.
D: Modified L-17Cs for Improved Multi-Arm DLVA Performance
In many DLVA applications, minimum baseline recovery-time and pulse transit-time are crucial. Successive-
detection DLVAs (SDLVAs) are often used in such instances. However, SDLVAs have much inferior
linearity, temperature-behavior, and general flexibility, compared to what is obtainable using the L-17C.
Further, SDLVAs are usually very expensive. Consequently, in order to approximate the best combined
features of the L-17C and the SDLVA, ANADYNE has developed a line of modified L-17Cs intended to be
used in two-arm and three-arm DLVAs.
For many multi-arm DLVAs, a dynamic range (power) of only 25 to 35 dB in each arm is needed. It can then
be very helpful to cut out the A3 linear gain stage (which also inactivates L1 and L2, the first two log stages)
in at least the high-level arms. This can have several advantages:
1.
Since the linear gain preceding the remaining log stages is reduced by a factor of 16, the
DLVA is much less sensitive to DC shifts caused by temperature or other factors.
2.
The noise contribution of the high-level arm to the overall DLVA noise is greatly reduced.
This can be important when the low-level arm is in a high-gain (low signal-level) state.
3.
The transit delay through the cut L-17C is reduced to about 10 to 12 ns. This delay can be
made fairly constant over the dynamic range by adjustment of the A2 speedup net.
4.
For each arm in which the A3 stage has been cut, the baseline recovery time can be
substantially reduced. This is expected because of the reduced dynamic range. For
example, using the HEROTEK DT2018 (2 to 18 GHz) coaxial tunnel-diode detector we have
achieved a 100 ns recovery time (to within ~1 dB of the baseline, using 5 recovery nets)
over a dynamic (logging) range of 0 to -42 dBm. Killing L1, the first log stage (by grounding
pin 9), reduces the start of logging by 6 dB, to -36 dBm. This decreases the recovery time
to about 50 ns, using the same recovery nets. (See Figs. 10(d) and 10(f) in the application
notes.)
By cutting A3, (which eliminates the contributions of both L1 and L2, and thus reduces the start of logging by
12 dB), even greater reductions in recovery time may be obtained. Thus, L-17C multi-arm DLVAs can have
recovery properties which are quite competitive with those of the more expensive (and less accurate)
SDLVAs. Of course, the penalty for lowering the start of logging by 12 dB is that 12 dB of additional RF gain
must be used to maintain the start of logging at the same RF input level.
5.
A given level of recovery can be achieved with fewer recovery nets, and the sensitivity of
the recovery time to detector waveform variations is reduced.
Consequently, ANADYNE now makes available a version of the L-17C in which the signal connection to A3
has been laser-cut. This short- dynamic-range version of the L-17C is designated the L-17C-2. At the
present time, the L-17C-2 is being made available at the same price as the normal L-17C, in both pinout #41
 |
 |
 |
 |
 |