Testbench - Canary 12 Channel Mixer
Test Report on: Canary 12 Channel Mixer
Date: May 1975, £496.00 EX-VAT.
The Canary 12 channel mixer is an impressive looking unit with an appearance very much like a small studio console. The 12 input channels each have an input gain control, a master slider, bass, middle, treble, echo send and 'on stage monitor' (fold back) level controls. There is also a small button, which puts any single channel signal on the monitor headphones, and a separate level meter for each channel.
There are three mixing stages: one to drive the output line; one for the echo send and the third for the 'on stage monitor'. Each had a master fader; and the output line and 'on stage monitor' mixers also have good sized output meters. The headphone monitor can be switched to the output line, the 'on stage' monitor on to P. F. L. (pre fade listen).
The case construction is of wood, with solid teak ends, and vinyl covered top, back, bottom and front edge. This supports a large blue painted metal front panel which is covered from end to end in controls and meters which makes the unit attractive.
Most of the electronic construction is on one huge glass fibre printed circuit board which covers almost the whole area of the front panel. The reverb electronics and power supplies are mounted in the bottom of the case.
The standard of soldering and quality of components is good and there are many expensive components. The fitting of many of the components is bad however. The circuit board is grossly overcrowded in some areas and empty in others. There is insufficient clearance between the backs of the faders and the input amplifiers which resulted in an intermittent fault on channel 10 when the front panel was touched. This was cured by bending the transistors 'out of the way'. In the first few minutes of testing the headphone monitor developed a fault. No attempt was made to find the cause but it could also be due to the poor mounting of components.
I am going to deviate from the normal method of tabulating results because most figures require explanation.
This is the maximum signal level which can be fed into the 'hi' input on any channel, with the gain control set at 'max' before overloading occurs. The figures for the 12 channels varied from 14mV to 17mV RMS; or expressed in dBm that is -35dBm to -33dBm, and an average of -34dBm (15mV RMS).
The input level meters indicated from -0.4 volume units to —2 volume units. Even with maximum input the meters are insensitive! The meter on channel 6 did not work at all.
If all faders are set to zero except for one channel at a time, then the noise at the output is divided by the gain of the system, the result is the equivalent noise referred to the input. This was done with channels one to four and the results were an average of -100dBm. as outlined in the table:
|Ch.||Gain||OP Noise||IP Noise|
This is the ratio of maximum input before overloading compared to the noise level. For this mixer the maximum input is —34dBm and the noise is -100dBm so the dynamic range is 66dB.
The manufacturers quote the sensitivity as -45dBm. If we compare our -100dBm input noise with this level we get a signal to noise ratio of 55dB for one channel. This is very much worse than the 70dB quoted by the makers. Their figure is impossible if the entire dynamic range is only 66dB.
Measured at 1 KHz with an input signal of -45dBm, the input gain and fader at max. and the line output set to a 0dBm level. Channels 1 to 4 were measured and they gave results of 0.8%, 0.66%, 0.62% and 0.65%.
Almost all this distortion was second harmonic and the figures are far worse than the less than 0.1% quoted by the makers for conditions which are far less favourable than those quoted above.
I was able to get slightly better figures by setting all the gain controls to maximum on one channel and all the others to zero. The input level was then reduced to give an output level of 0dBm. This gave T.H.D. figures of 0.32%, 0.28%, 0.32%.
Noise from the mixing stage should be negligible but with all faders set to zero, except the line faders, an output noise level of —58.5dB was monitored.
All the above noise and distortion measurements were made with all unused faders set to zero, tone controls set to midway, and a 10K dummy source in the 'Hi' input for noise measurement only.
|BASS @ 30Hz||MIDDLE @ 500Hz||TREBLE @ 10KHz|
A square wave test which was to set the frequency response on flat if possible (i.e. set the tone controls so that a square wave in will give a square wave out) revealed treble boost which could not be compensated with the controls. Consequently it was decided not to measure the frequency response.
The appearance is so good that I was disappointed to find that the performance was not all it should be.
The only makers' specification figures I could confirm were the ranges of the tone controls.
Most of the troubles seem to be related to driving a meter from each of the input channel amplifiers without having a proper meter driving circuit. The meters are not sensitive enough anyway.
It arrived with one meter not working and a fault on channel 10 and the headphone amplifier developed a fault in the first few minutes of testing.