Microphone
input measurements:
Attenuator
on mike input, using signal generator (1kHz), increasing input voltage until
output one becomes distorted. Measured at different manually set mike gains (1
to 16) [in MENU]. Voltages are peak values.
Mic
gain |
Max.
mic input [mV] |
No
of LEDs |
Out
voltage [V] |
Gain |
1 |
59 |
15 |
0.56 |
9.5 |
2 |
60 |
18 |
0.90 |
15 |
3 |
60 |
20 |
1.3 |
21.7 |
4 |
39 |
20 |
1.3 |
33.3 |
5 |
26 |
20 |
1.3 |
50 |
6 |
18 |
20 |
1.3 |
72.2 |
7 |
12 |
20 |
1.3 |
108 |
8 |
8.2 |
20 |
1.3 |
159 |
9 |
5.6 |
20 |
1.3 |
232 |
10 |
3.9 |
20 |
1.3 |
333 |
11 |
2.8 |
20 |
1.3 |
464 |
12 |
2.06 |
20 |
1.3 |
631 |
13 |
1.51 |
20 |
1.3 |
861 |
14 |
1.16 |
20 |
1.3 |
1121 |
15 |
0.95 |
20 |
1.3 |
1368 |
16 |
0.80 |
20 |
1.3 |
1625 |
In auto-gain mode, the gain depends on input signal. There are some measurements*:
mike input |
Out voltage [mV] |
Gain |
0.5 |
400 |
800 |
0.7 |
560 |
800 |
1.0 |
800 |
800 |
1.2 |
910 |
758 |
1.5 |
960 |
640 |
2.0 |
1010 |
505 |
3.0 |
1030 |
343 |
4.0 |
1050 |
263 |
5.0 |
1060 |
212 |
10.0 |
1100 |
110 |
When signal is relatively weak, the AGC uses the largest gain, which is 800 (58dB). It approximately corresponds to the manual gain at level 13.
Frequency
characteristics (mike input)
Mike
gain is set manually at position 3 [in MENU]:
Frequency
[kHz] |
Input
signal [mV] |
Output
signal [mV] |
Gain
[dB] |
0.01 |
3.5 |
3 |
-1.3 |
0.02 |
6 |
30 |
14.0 |
0.05 |
8.5 |
130 |
23.7 |
0.1 |
9.5 |
195 |
26.2 |
0.2 |
10 |
220 |
26.8 |
0.5 |
10 |
220 |
26.8 |
1.0 |
10 |
220 |
26.8 |
2.0 |
10 |
215 |
26.6 |
5.0 |
10 |
205 |
26.2 |
10.0 |
10 |
180 |
25.1 |
15.0 |
10 |
145 |
23.2 |
20.0 |
10 |
125 |
21.9 |
Noise
is approx. 2mV peak (at the output - measured with oscilloscope); 50Hz component
is the most noticeable.
Frequency
characteristics of line input:
Flat from 50Hz to 20kHz; at 20Hz gain is –0.4dB, at 10Hz gain is –1.5dB. Noise is about 10/32000 (voltage), so about –70dB.
Frequency characteristics graphically (mic in: blue broken line; line in: green solid line):
Noise
measurements:
1. Mike input
There were two resistors (1K) connected to the TRV900E mike input (between signal and GND). Manual gain setting has been used. Gain slider changed from value 1 to 16. The obtained average RMS (in dB) on the left (L) and on the right (R) channel were the following:
Mic
gain |
Noise
L [dB] |
Noise R [dB] |
1 |
-62.3 |
-62.4 |
2 |
-62.7 |
-62.8 |
3 |
-62.5 |
-62.6 |
4 |
-62.6 |
-62.6 |
5 |
-62.8 |
-62.8 |
6 |
-62.5 |
-62.4 |
7 |
-62.5 |
-62.4 |
8 |
-61.6 |
-61.7 |
9 |
-60.5 |
-61.0 |
10 |
-59.0 |
-59.8 |
11 |
-56.7 |
-58.2 |
12 |
-55.2 |
-56.7 |
13 |
-52.9 |
-54.7 |
14 |
-51.9 |
-53.5 |
15 |
-50.7 |
-52.2 |
16 |
-50.2 |
-51.2 |
In auto-gain (AGC), the obtained average RMSs are -59.3dB for left and right channel.
However, noise characteristics over frequencies is much more interesting [pictures]. Namely, when using manual gain, there are some spikes existing at frequencies 50Hz, 100Hz, 150Hz, 200Hz, ... (note that I'm using PAL model), while in auto-gain (AGC), there is only one spike at 150Hz. Taking into account that signal amplification of the AGC is approximately the same to the manual one at level 13, it is clear that AGC gives much better signal/noise (S/N) ratio than when using manual gain (comparing RMS values, AGC gives about 4 to 6 dB better S/N ratio; compare also figures 'Mic gain 13' and 'Mic AGC'). Funny, isn't it? It seems that this low-frequency "contamination" comes from the manual-gain control (electronic volume M5222 and low-noise amp 4572) [picture]. I'll try to figure out why does it happen...
Noise is distincly weaker when using line in input (in VTR mode). The average RMSs are -78.0dB for left and -77.9dB for right channel. The frequency characteristic is given here. It is interesting that I've got a quite better result than John on VX2000.
Is it possible to get better S/N ratio at mike input? Read this.
How to build your own elastic suspension ("shockmount") and mike preamp with attenuator for AKG CK91 capsule? Read this.