Attenuation | Amplitude (mV) | Noise (mV) | ratio | Amplitude | Noise | ratio |
0 db | 420 | 20 | 26.4 db | 90 | 26 | 11 db |
-10 db | 136 | 8 | 24.6 db | 25 | 20 | 2 db |
-20 db | 6 | 21 | -11 db | |||
-30 db | 14 | 6 | 7.4 db |
The most usefull way to compare
them would be to compare the outputs of a postamp/comparator after it is
optimised for use with each preamp. Pulse width and rising edge jitter
would be compared over a wide signal input range. Since the output pulse
width can be used as an indicator of input pulse width, the postamp needs
to operate in a linear region and not shape the pulse more than the VTX
preamp already does.
To accomplish this, it was
decided to modify the MEGA postamp/comparator which was available
and also under consideration for use in the experiment
Initial testing showed
that the IC used as the postamp has very nonlinear response when operated
at the lower gain required, even without the shaping R-C networks, which
leads to reducing the input pulse to about half it's width. This behavior
is documented in the data sheets for this part, the LM592. To solve this
problem a pin-compatable IC, the uA733/LM733, was used instead.
Again, comparing signal
to noise ratios after the optimised postamp at the input to the comparator
is a usefull indicator of how things are going. It is worth noting that
the input to this comparator is differential and that the postamp has converted
the single ended VTX signal to a differential one at this point.
Attenuation | Amplitude (mV) | Noise (mV) | ratio | Amplitude | Noise | ratio |
0 db | 1,500 | 50 | 29.5 db | 920 | 50 | 25.2 db |
-10 db | 560 | 25 | 27 db | |||
-20 db | 175 | 25 | 17 db | 90 | 40 | 7 db |
-30 db | 56 | 26 | 6.4 db | 32 | 42 | -2.4 db |
This leads to the following
comparator outputs.
Attenuation | Width | Jitter | Width | Jitter | comment |
0 db | 40 | <1 | ~56 | 2 | (14 falling edge) |
-10 db | 30.8 | 1 | |||
-20 db | 22 | 1 | 24 | 4 | (double pulsing) |
-30 db | 11 | 2 | 13 | - | (double pulsing) |
Some of the double pulsing on the MEGA preamp setup is due to a rising tail on the comparator input indicating that the shaping is not ideal.
Conclusion
From these tests it appears that the
VTX preamp is a clear winner, mainly due to much lower noise and higher
output which requires less postamplification. The only disadvantage would
be if the single ended drive leads to more noise pickup over the cable
lengths required. This problem could be overcome by adding a differential
twisted pair driver IC, such as the Elantec EL2140, near the detector end
and doubling the post amp gain to make up for loss in the cable terminations.
List of Figures
fig. 1
VTX preamp out
0 db
envelope mode
fig. 2
" "
" - 10 db
"
fig. 3
" "
" - 30 db
"
fig. 4
MEGA "
"
0 db
"
fig. 5
" "
" - 10 db
"
fig. 6
" "
" - 20 db
"
fig. 7
" "
" - 30 db
"
fig. 8
NE592 frequency response data sheet
fig. 9
uA733 "
" "
"
fig. 10
MEGA postamp modifications
fig. 11
VTX preamp to MEGA postamp
0 db average mode
fig. 12
"
0 db envelope mode
fig. 13
"
+ 5 db preamp overdrive
fig. 14
"
- 10 db average
fig. 15
"
- 10 db envelope
fig. 16
"
- 20 db "
fig. 17
"
- 30 db "
fig. 18
MEGA preamp to MEGA postamp 0 db
"
fig. 19
"
- 20 db "
fig. 20
"
- 30 db average
fig. 21
"
- 30 db envelope