Hello to everyone.
This is a review and test of the Bel Canto PRe6 Gen II multichannel preamplifier. The original PRe6 was released in 2003 and cost about US$3,990 at the time. As I understand, the Gen II version incorporates different components. As the printed circuit board bears the date of release of the first version and no bodges or unused solder pad are visible, I think there is no major revisions between the two versions. The firmware version briefly displayed on screen a few seconds after leaving standby is dated from 2003. So, I do not expect software mod either.
Part I: Presentation
The Bel Canto PRe6 is an 8 channels pure analogue preamplifier that can be configured in many ways by software control. The back of the unit shows it all:
Above, a row of 8 XLR outputs and a pair of XLR inputs for stereo. Bottom, 8 RCA output sockets, 2 others for stereo tape output and 2 for input tape monitoring and finally 10 pairs of RCA input sockets. There are two groups of multichannel RCA inputs and each can be independently configured as either an 8 channels input (7.1), or a 6 channels input (5.1) and an additional stereo input or as 4 stereo inputs. The 8 analogue outputs, whether XLRs or RCAs, follows the input configuration. When 5.1 plus stereo or 4 stereo channels is selected, the two rear surround channels become an independent "zone 2" stereo output with independent volume control. This is a very versatile preamplifier!
You may have noticed that each RCA input is accompanied by a pair of LEDs of different colors. They lit up when the associated input is selected and also indicate how the multichannel inputs are configured to help identifying where plugging cables. But as it is inadvisable to plug and unplug cables while the preamplifier is live (and it is always live, even in standby), this nice touch is of limited value in use.
There are also a pair of trigger outputs for 3.5 mm jacks and an RS232 interface on Sub-D connector for system integration.
This preamplifier is massive. It is a bit larger than usual and weights a significant 11 kg (24lbs). The brushed aluminum front fascia is impressively finished. Unfortunately, I find that the front panel blue dot matrix display looks a bit cheap. It cannot shows many characters. Nothing comparable to the better looking and more useful display of the Linn Exotik I recently reviewed. More annoying, the Bel Canto falls short of expectation regarding the quality of the RCA input sockets. You can judge for yourself:
Some sockets are not perfectly well in line with the holes in the back panels and are pressed and crooked. I do not now if this is the factory condition or if it happened by accident, but I have to report that another group of 4 input sockets more to the left are in the same conditions... on an almost US$4,000 device, this is unfortunate. Again, nothing comparable to the quality and robustness of the RCA connectivity of said Linn preamp.
This is regrettable, because when we pick a look inside and put the overall mechanical built quality under scrutiny, the PRe6 more than redeems itself:
The quality of the metal work is of a high standard, with thick and very well coated plates. All chassis screws are mounted through inserts and have locking washers.
The main circuit board is a four layers job and is neat and clean. The entire 8 channels audio circuit lies on the top right corner on the above picture, a good illustration of the compactness that could be achieved with highly integrated electronics already twenty years ago. The signal path is straightforward. The stereo XLR inputs are unbalanced by a Burr Brown INA2134 integrated dual differential input receivers. Then follows an impressive array of sealed relays for input configuration and switching. Then, each of the eight available input channels is buffered by a socketed Burr Brown OPA134 single op-amp. The volume control is implemented through 4 Burr Brown PGA2310 stereo integrated circuits. Last, 8 Texas Instruments THS4130 fully differential amplifiers (FDAs) convert the unbalanced signals out of the volume control ICs in differential signals for the XLR outputs. The manufacturer claims that the balanced outputs are capable of driving cables up to 152 m (500 ft) of length. I was not able to find if the unbalanced outputs are taken directly at the outputs of the PGA2310s (which are rated to be able to drive 600 ohms loads and up to a hefty 1,000 pF of capacitance) or if they are taken at one of the two outputs of each THS4130. All balanced and unbalanced outputs can be muted by output sealed relays. Last, the tape outputs are buffered by a Burr Brown OPA2134 dual op amps. There is no power switch, nor real standby mode. Actually, entering standby just hard mutes the output by relays and disables the front dot matrix display screen and rear LEDs, and that's it. All the rest remains powered.
Even the top cover has been given some thoughts, but I admit I cannot figure out the purpose:
An elegant universal remote control unit built in aluminum and with a touch screen LCD display was provided. Sadly, the screen has partially faded away, but it remains fully functional:
The PRe6 can be configured and operated entirely with the seven metal front push buttons. Their tactile response is not as soft and pleasant than those of the Linn, again. It is possible to customize the name display for each input, set volume offset between channels and inputs, the level of a soft mute and, above all, set the unity gain point anywhere in a 25 dB range. That means that it is possible to adjust the total amount of gain of the preamplifier to the sources output levels and/or the input sensitivity of power amplifiers while still enjoying the full adjustment range of the volume control. The volume control itself is incremented in unity steps from 0 to 120, which are equal to as much 0.5 dB steps, i.e. a range of adjustment of 60 dB, more than enough to cope even with the wildest compressed or uncompressed music program materials. Left/right balance adjustment is only accessible via the remote control unit.
Part II: Measurements
All measurements were taken with an Audio Precision System One+DSP SYS222A (nominal input impedance: 100 kohms per phase in parallel with 170 pF). Unless otherwise noted, all measurements were performed using standard Audio Precision tests. The Audio Precision has always been given the 30 minutes preconditioning period mandated in its calibration procedure.
The PRe6 user manual's states that "sonic performance will start to stabilize after approximately 40 hours of continuous power up". So I repeated some measurements after such a preconditioning period, but found no differences. Save for few exceptions, there were no differences between the unbalanced and balanced signal paths either. That means that the overall level of performance is set by the internal unbalanced signal path and that the balanced inputs and outputs are transparent. By the way, the balanced inputs and outputs gain is compensated to match the level of the unbalanced inputs and outputs. So, I will not bother to show all measurements. All the below measurements are thus taken from the unbalanced inputs and outputs unless otherwise noted. Last important point: the PRe6 was measured in its default factory settings, including the unity gain setting point. When I took measurements at other gain settings, the results were all identical (except of course input or output clipping points) to those taken with the factory settings. This consistency is to the credit of the PRe6. Overall, this preamplifier is at its best at unity gain. It was obviously designed to be used in a contemporary environment of high level output digital sources.
Let's first see a dashboard at unity gain, the closer to 2 V RMS in/2 V RMS out that is possible, to follow the practice set by @amirm. Please keep in mind that it is not possible to get at the output exactly the same voltage than at the input because the volume control is stepped:
The dashboard bar-graphs shows the least good of the two front channels. There is some very low-level hum and very low harmonic distortion, with a somewhat unfortunate spectrum of regularly decreasing odd order harmonics, though. Nevertheless, the SINAD is much better than CD-level quality. The levels of the two channels are matched within 0.08 dB (not shown), which is very good.
Reducing the level at 500 mV RMS (a widely followed standard input level for measurement in the old days) shows less distortion, but now even order harmonic products make an appearance:
Take note that the displayed FFT and the bar-graph THD+N are referred to the analogue output level, which is now only 500 mV RMS. That's why they appears to be much less good than on the previous dashboard, where all is referred to about 2 V RMS (about 12 dB higher level than 500 mV RMS).
To compare with the first dashboard, here is another one taken from the XLR stereo inputs to the balanced outputs at a standard 4 V RMS level:
The improvement is minimal.
04/20/2026 edition: The stereo balanced input common mode rejection ratio (CMRR) is about 67 dB over a 50 kHz bandwidth, degrading just slightly between 10 kHz and 50 kHz from 66.6 to 65.4 dB in the right channel. These are relatively good results, but not up to the typical performance specified for the INA2134 dual differential input receivers.
The THD+N amplitude versus input level amplitude sweep in absolute value (0 dBV=1 V RMS) shows a clipping level at about 7.9 V RMS output:
The SMPTE intermodulation (low frequency+high frequency tones) vs level sweep shows good results:
So does the CCIF intermodulation (twin high frequency tones) vs level sweep, revealing good high frequency linearity:
Finally, the last intermodulation test shows transient intermodulation distortion (DIM standard: high frequency sine + filtered square wave) vs level:
Again, this seems to be a good result, largely noise dominated up to +6 dBV of input/output level signal.
The downward slope above 5 kHz of the curves showing THD+N vs frequency sweep at a standard 500 mV input level in a 20 kHz measurement bandwidth implies that the distortion spectrum contains harmonics higher than rank 4 (the 4th harmonic of input signal higher than 5 kHz are above the 20 kHz limit of the measurement bandwidth), which is consistent with the dashboards:
At 2 V RMS (the level at which @amirm performs this test with unbalanced devices, I believe) and inside an 80 kHz measurement bandwidth, we see frequency-dependent effects:
The balanced outputs are a bit better on this test, presumably thanks to their higher output level voltage swing:
I have measured signal to noise ratio at unity gain, 2 V RMS in, and at the maximum input/output level of 8.5 V RMS (at less than 0.1% THD+N). I also followed a old habit of German reviewers to measure signal to noise ratio (always in a 22 Hz to 20 kHz bandwidth) referred to usual output levels set by the volume control during casual listening with a standard 500 mV RMS input level signal. Overall, the results are quite good. First, from unbalanced inputs to unbalanced outputs:
Second, from unbalanced inputs to balanced outputs:
The separation between the two front channels is good, rather than great. We are far from the state of the art:
It improves a bit when measured from the balanced outputs:
The frequency response is almost flat all the way up to 200 kHz:
This is a very wide bandwidth preamplifier! As a result, the input to output phase response is almost flat over the entire audio band:
Audio Precision provides a test of the input/output linearity, in order to check if the output level does change linearly with the change of the input level. The PRe6 is superbly linear, save for some noise effects, albeit small, under about -95 dBV:
This curves can be compared to that of the Vincent SAV-C1, in which a noise-gate function impacts very low level linearity.
Finally, one of the most important features of a preamplifier is, to my mind, the accuracy of the volume control, which has to preserve an identical balance between the left and right channels (or all channels of a multichannel preamplifier) whatever the volume setting. I have created a test to show both the accuracy of the volume control step size and the matching of two channels. Unfortunately, here I encountered a quirk: the Audio Precision right channels was unable to settle to any reading under somewhat -20 dB from maximum volume. I suspect this is due to some high frequency noise on that channel which disturbs the steadiness of the signal ever so slightly. Further investigation are needed. Accordingly, I measured the tracking of two more widely separated channels instead:
There is nothing to report but superb tracking and step size accuracy.
Lastly, for the fun, I decided to perform a test with my killer test signal generator to separate the men from the boys: an old Nokia 3310 GSM cell phone making a call placed, with suitable orientation to get the greatest radiated emission, right near the input RCA sockets under scrutiny. I like to hear no demodulation of the 217 Hz pulse frequency of GSM packets in the loudspeaker as a testament to the care taken by the designer to reduce the RFI sensitivity of his device. This test was inspired by reading a note from Kymberly Schmidt, Maxim Integrated. Here is an FFT of a 1 kHz test signal at 2 V RMS input, unity gain, in that condition:
All effects due to the demodulation of the RFI radiated by the cell phone appears to remain at levels which should be inaudible, especially at usual volume settings for casual listening, which is good.
CONCLUSIONS:
The PRe6 Gen II is versatile, well built, good performing multichannel preamplifier. However, a few minor manufacturing details could have been improved. One of the strength of this preamp is its wide array of available inputs, including two multichannel ones, which is relatively rare. It seems a bit bulky to me, but that's a matter of taste.
This ends the review. I hope you find it interesting.
This is a review and test of the Bel Canto PRe6 Gen II multichannel preamplifier. The original PRe6 was released in 2003 and cost about US$3,990 at the time. As I understand, the Gen II version incorporates different components. As the printed circuit board bears the date of release of the first version and no bodges or unused solder pad are visible, I think there is no major revisions between the two versions. The firmware version briefly displayed on screen a few seconds after leaving standby is dated from 2003. So, I do not expect software mod either.
Part I: Presentation
The Bel Canto PRe6 is an 8 channels pure analogue preamplifier that can be configured in many ways by software control. The back of the unit shows it all:
Above, a row of 8 XLR outputs and a pair of XLR inputs for stereo. Bottom, 8 RCA output sockets, 2 others for stereo tape output and 2 for input tape monitoring and finally 10 pairs of RCA input sockets. There are two groups of multichannel RCA inputs and each can be independently configured as either an 8 channels input (7.1), or a 6 channels input (5.1) and an additional stereo input or as 4 stereo inputs. The 8 analogue outputs, whether XLRs or RCAs, follows the input configuration. When 5.1 plus stereo or 4 stereo channels is selected, the two rear surround channels become an independent "zone 2" stereo output with independent volume control. This is a very versatile preamplifier!
You may have noticed that each RCA input is accompanied by a pair of LEDs of different colors. They lit up when the associated input is selected and also indicate how the multichannel inputs are configured to help identifying where plugging cables. But as it is inadvisable to plug and unplug cables while the preamplifier is live (and it is always live, even in standby), this nice touch is of limited value in use.
There are also a pair of trigger outputs for 3.5 mm jacks and an RS232 interface on Sub-D connector for system integration.
This preamplifier is massive. It is a bit larger than usual and weights a significant 11 kg (24lbs). The brushed aluminum front fascia is impressively finished. Unfortunately, I find that the front panel blue dot matrix display looks a bit cheap. It cannot shows many characters. Nothing comparable to the better looking and more useful display of the Linn Exotik I recently reviewed. More annoying, the Bel Canto falls short of expectation regarding the quality of the RCA input sockets. You can judge for yourself:
Some sockets are not perfectly well in line with the holes in the back panels and are pressed and crooked. I do not now if this is the factory condition or if it happened by accident, but I have to report that another group of 4 input sockets more to the left are in the same conditions... on an almost US$4,000 device, this is unfortunate. Again, nothing comparable to the quality and robustness of the RCA connectivity of said Linn preamp.
This is regrettable, because when we pick a look inside and put the overall mechanical built quality under scrutiny, the PRe6 more than redeems itself:
The quality of the metal work is of a high standard, with thick and very well coated plates. All chassis screws are mounted through inserts and have locking washers.
The main circuit board is a four layers job and is neat and clean. The entire 8 channels audio circuit lies on the top right corner on the above picture, a good illustration of the compactness that could be achieved with highly integrated electronics already twenty years ago. The signal path is straightforward. The stereo XLR inputs are unbalanced by a Burr Brown INA2134 integrated dual differential input receivers. Then follows an impressive array of sealed relays for input configuration and switching. Then, each of the eight available input channels is buffered by a socketed Burr Brown OPA134 single op-amp. The volume control is implemented through 4 Burr Brown PGA2310 stereo integrated circuits. Last, 8 Texas Instruments THS4130 fully differential amplifiers (FDAs) convert the unbalanced signals out of the volume control ICs in differential signals for the XLR outputs. The manufacturer claims that the balanced outputs are capable of driving cables up to 152 m (500 ft) of length. I was not able to find if the unbalanced outputs are taken directly at the outputs of the PGA2310s (which are rated to be able to drive 600 ohms loads and up to a hefty 1,000 pF of capacitance) or if they are taken at one of the two outputs of each THS4130. All balanced and unbalanced outputs can be muted by output sealed relays. Last, the tape outputs are buffered by a Burr Brown OPA2134 dual op amps. There is no power switch, nor real standby mode. Actually, entering standby just hard mutes the output by relays and disables the front dot matrix display screen and rear LEDs, and that's it. All the rest remains powered.
Even the top cover has been given some thoughts, but I admit I cannot figure out the purpose:
An elegant universal remote control unit built in aluminum and with a touch screen LCD display was provided. Sadly, the screen has partially faded away, but it remains fully functional:
The PRe6 can be configured and operated entirely with the seven metal front push buttons. Their tactile response is not as soft and pleasant than those of the Linn, again. It is possible to customize the name display for each input, set volume offset between channels and inputs, the level of a soft mute and, above all, set the unity gain point anywhere in a 25 dB range. That means that it is possible to adjust the total amount of gain of the preamplifier to the sources output levels and/or the input sensitivity of power amplifiers while still enjoying the full adjustment range of the volume control. The volume control itself is incremented in unity steps from 0 to 120, which are equal to as much 0.5 dB steps, i.e. a range of adjustment of 60 dB, more than enough to cope even with the wildest compressed or uncompressed music program materials. Left/right balance adjustment is only accessible via the remote control unit.
Part II: Measurements
All measurements were taken with an Audio Precision System One+DSP SYS222A (nominal input impedance: 100 kohms per phase in parallel with 170 pF). Unless otherwise noted, all measurements were performed using standard Audio Precision tests. The Audio Precision has always been given the 30 minutes preconditioning period mandated in its calibration procedure.
The PRe6 user manual's states that "sonic performance will start to stabilize after approximately 40 hours of continuous power up". So I repeated some measurements after such a preconditioning period, but found no differences. Save for few exceptions, there were no differences between the unbalanced and balanced signal paths either. That means that the overall level of performance is set by the internal unbalanced signal path and that the balanced inputs and outputs are transparent. By the way, the balanced inputs and outputs gain is compensated to match the level of the unbalanced inputs and outputs. So, I will not bother to show all measurements. All the below measurements are thus taken from the unbalanced inputs and outputs unless otherwise noted. Last important point: the PRe6 was measured in its default factory settings, including the unity gain setting point. When I took measurements at other gain settings, the results were all identical (except of course input or output clipping points) to those taken with the factory settings. This consistency is to the credit of the PRe6. Overall, this preamplifier is at its best at unity gain. It was obviously designed to be used in a contemporary environment of high level output digital sources.
Let's first see a dashboard at unity gain, the closer to 2 V RMS in/2 V RMS out that is possible, to follow the practice set by @amirm. Please keep in mind that it is not possible to get at the output exactly the same voltage than at the input because the volume control is stepped:
The dashboard bar-graphs shows the least good of the two front channels. There is some very low-level hum and very low harmonic distortion, with a somewhat unfortunate spectrum of regularly decreasing odd order harmonics, though. Nevertheless, the SINAD is much better than CD-level quality. The levels of the two channels are matched within 0.08 dB (not shown), which is very good.
Reducing the level at 500 mV RMS (a widely followed standard input level for measurement in the old days) shows less distortion, but now even order harmonic products make an appearance:
Take note that the displayed FFT and the bar-graph THD+N are referred to the analogue output level, which is now only 500 mV RMS. That's why they appears to be much less good than on the previous dashboard, where all is referred to about 2 V RMS (about 12 dB higher level than 500 mV RMS).
To compare with the first dashboard, here is another one taken from the XLR stereo inputs to the balanced outputs at a standard 4 V RMS level:
The improvement is minimal.
04/20/2026 edition: The stereo balanced input common mode rejection ratio (CMRR) is about 67 dB over a 50 kHz bandwidth, degrading just slightly between 10 kHz and 50 kHz from 66.6 to 65.4 dB in the right channel. These are relatively good results, but not up to the typical performance specified for the INA2134 dual differential input receivers.
The THD+N amplitude versus input level amplitude sweep in absolute value (0 dBV=1 V RMS) shows a clipping level at about 7.9 V RMS output:
The SMPTE intermodulation (low frequency+high frequency tones) vs level sweep shows good results:
So does the CCIF intermodulation (twin high frequency tones) vs level sweep, revealing good high frequency linearity:
Finally, the last intermodulation test shows transient intermodulation distortion (DIM standard: high frequency sine + filtered square wave) vs level:
Again, this seems to be a good result, largely noise dominated up to +6 dBV of input/output level signal.
The downward slope above 5 kHz of the curves showing THD+N vs frequency sweep at a standard 500 mV input level in a 20 kHz measurement bandwidth implies that the distortion spectrum contains harmonics higher than rank 4 (the 4th harmonic of input signal higher than 5 kHz are above the 20 kHz limit of the measurement bandwidth), which is consistent with the dashboards:
At 2 V RMS (the level at which @amirm performs this test with unbalanced devices, I believe) and inside an 80 kHz measurement bandwidth, we see frequency-dependent effects:
The balanced outputs are a bit better on this test, presumably thanks to their higher output level voltage swing:
I have measured signal to noise ratio at unity gain, 2 V RMS in, and at the maximum input/output level of 8.5 V RMS (at less than 0.1% THD+N). I also followed a old habit of German reviewers to measure signal to noise ratio (always in a 22 Hz to 20 kHz bandwidth) referred to usual output levels set by the volume control during casual listening with a standard 500 mV RMS input level signal. Overall, the results are quite good. First, from unbalanced inputs to unbalanced outputs:
| Input Level | Output Level | Left Channel SNR (ref. to output level) | Right Channel SNR (ref. to output level) |
|---|---|---|---|
| 500 mV | 30 mV | 68.6 dB (71 dBA) | 68.8 dB (70.9 dBA) |
| 500 mV | 300 mV | 88.2 dB (90.2 dBA) | 88.3 dB (90.4 dBA) |
| 500 mV | 1V | 92.8 dB (95.1 dBA) | 92.8 dB (94.9 dBA) |
| 2 V | 2 V | 105.3 dB (107.7 dBA) | 105.3 dB (107.4 dBA) |
| 8.5 V | 8.5 V | 118 dB (120 dBA) | 117.9 dB (120 dBA) |
Second, from unbalanced inputs to balanced outputs:
| Input Level | Output Level | Left Channel SNR (ref. to output level) | Right Channel SNR (ref. to output level) |
|---|---|---|---|
| 500 mV | 30 mV | 68.8 dB (71 dBA) | 68.8 dB (71.1 dBA) |
| 500 mV | 300 mV | 88.5 dB (90.5 dBA) | 88.3 dB (90.4 dBA) |
| 500 mV | 1V | 93 dB (95.2 dBA) | 92.5 dB (95.2 dBA) |
| 2 V | 2 V | 105.2 dB (107.2 dBA) | 104.6 dB (107 dBA) |
| 8.5 V | 8.5 V | 118.1 dB (119.9 dBA) | 117.5 dB (119.9 dBA) |
The separation between the two front channels is good, rather than great. We are far from the state of the art:
It improves a bit when measured from the balanced outputs:
The frequency response is almost flat all the way up to 200 kHz:
This is a very wide bandwidth preamplifier! As a result, the input to output phase response is almost flat over the entire audio band:
Audio Precision provides a test of the input/output linearity, in order to check if the output level does change linearly with the change of the input level. The PRe6 is superbly linear, save for some noise effects, albeit small, under about -95 dBV:
This curves can be compared to that of the Vincent SAV-C1, in which a noise-gate function impacts very low level linearity.
Finally, one of the most important features of a preamplifier is, to my mind, the accuracy of the volume control, which has to preserve an identical balance between the left and right channels (or all channels of a multichannel preamplifier) whatever the volume setting. I have created a test to show both the accuracy of the volume control step size and the matching of two channels. Unfortunately, here I encountered a quirk: the Audio Precision right channels was unable to settle to any reading under somewhat -20 dB from maximum volume. I suspect this is due to some high frequency noise on that channel which disturbs the steadiness of the signal ever so slightly. Further investigation are needed. Accordingly, I measured the tracking of two more widely separated channels instead:
There is nothing to report but superb tracking and step size accuracy.
Lastly, for the fun, I decided to perform a test with my killer test signal generator to separate the men from the boys: an old Nokia 3310 GSM cell phone making a call placed, with suitable orientation to get the greatest radiated emission, right near the input RCA sockets under scrutiny. I like to hear no demodulation of the 217 Hz pulse frequency of GSM packets in the loudspeaker as a testament to the care taken by the designer to reduce the RFI sensitivity of his device. This test was inspired by reading a note from Kymberly Schmidt, Maxim Integrated. Here is an FFT of a 1 kHz test signal at 2 V RMS input, unity gain, in that condition:
All effects due to the demodulation of the RFI radiated by the cell phone appears to remain at levels which should be inaudible, especially at usual volume settings for casual listening, which is good.
CONCLUSIONS:
The PRe6 Gen II is versatile, well built, good performing multichannel preamplifier. However, a few minor manufacturing details could have been improved. One of the strength of this preamp is its wide array of available inputs, including two multichannel ones, which is relatively rare. It seems a bit bulky to me, but that's a matter of taste.
This ends the review. I hope you find it interesting.
Last edited:
