I share with this post a review of the vintage Threshold FET Ten/e High Level preamplifier. This is a thirty years old model, built in the late 1980s by the legendary Nelson Pass and considered a reference by many sources.
I tried to discover why, via measurements in different working conditions and listening tests.
The case of the FET Ten/e is well built, compact and robust. It has:
- Stepped Nobles balance and volume controls;
- Tape outputs (and related selectors);
- Selector switch for five inputs (no phono stage; the PC unit was sold separately);
- Attenuator of -20dB (no amplification) and Mono switch.
It has also one XLR input and both XLR and RCA outputs, presents only in the “e” version.
All FET Ten/e units are not equally; this unit should be the last version with military grade parts, only bettered by the extremely rare FET 10/t. Serial number is 064xx.
Its intern is amazing: all discrete components, separated board gain modules and an attention to all details. It uses discrete J-FET and MOSFET.
Power supply (15V) is made by an over dimensioned outboard (not shown).
In ’90, the price in Europe was about 3200€.
SPECIFICATIONS
Frequency response: 1.5Hz–150kHz +0, –3dB
THD: 0.01% at 1V out.
Max output: 20V
Gain: 20dB max
Output impedance: 100 ohms.
MEASUREMENTS
All following measurements have been made with:
- Input: XLR; from digital to analog with resolution 24/192
- Output: XLR; from analog to digital with resolution 24/192
- Load: 9 Kohm
FREQUENCY RESPONSE
The frequency response is excellent, flat far beyond the audible band, with a difference a little difference of 0.15dB between channels.
In the phase shift we see the same excellence: distortion is 0 also at 10Hz (wow), despite the low load, 9KOhm. This means a coherence for bass. Less than 3 degrees at 20KHz, very good.
CROSSTALK
Crosstalk is very good, even if there is an asymmetry of about 15dB among channels, maybe due to circuit layout. At 1KHz we have a very good -85dB (on L) and an optimum -99dB (on R).
Curiously, increasing the input of 20dB to reach 1Vrms (Gain still at 20dB, so output reaches 10Vrms) only the lower curve (L over R) shifts down, increasing the gap to 18dB.
Switching ON the -20dB Attenuator, the curves became closer (5dB), but both shift up, to higher values.
SINGLE TONE DISTORTION
At 1Vrms input, 0dB Gain, SNR is a good 83dB, with a TD+N of -86dB, for both channels. Harmonic distortion is mainly in the second harmonic (-100dB) and decreases with others (-111dB, -124db …), inaudible. Few the differences between the channels. Excellent the noise to power supply, practically absent, below 120dB.
Maintaining fixed the output level at 1 Vrms and changing the Gain at 20dB and input level, SNR became 77.7dB, and TD+N of -79.6dB, for both channels. Harmonic distortion is due all to the second harmonic at -106dB, excellent. Unchanged the hum, always excellent.
With 1Vrms input and output but with -20dB Attenuator set to ON (Gain 0dB, Volume at 100%), the result is still very good: SNR is a good 83.2dB, with a TD+N of -86.7dB, for both channels. Harmonic distortion remains mainly in the second harmonic (-99dB) and decreases with others (-111dB, -124db …), inaudible. Few the differences between the channels; Unchanged the hum.
HARMONIC DISTORTION
At 1Vrms in the output, 0dB Gain (50% of volume) THD is 0.005% practically constant for all frequencies, below a very good 0.001% (-99dB), all due to second harmonic; the third one is about 0.0003% and it increases for values of frequency outside the audible band. The other harmonics are on lower values 0.00005%. THD+N is a good 0.005%. Excellent.
The graph (not shown) has not meaningful changes with 5Vrms input and Gain -14db (with or without attenuator).
In the standard condition, both THD and THD+N remain flat for all frequencies: about 0.0006% for THD (-104dB) and 0.01% (-80dB) for THD+N, excellent values. Distortion is mainly on the second harmonic; the others are practically inexistent.
The distortion for level is also very good, with optimal values around 0.2Vrms output; above, increases practically only the second harmonic.
Setting the Gain to max, and leaving the 1V as max output level, we have still very good results, with optimal values around 0.5Vrms output:
INTERMODULATION
Very good also the CCIF test: IMD is -94.5 dB. Symmetrical nonlinearities (odd-order), near input signals, are about -104dB; asymmetrical, even-order distortion at 1KHz is -99dB; TD+N is -83dB. Negligible differences (not shown) between channels.
With Gain at 20dB, 1 Vrms output, the numbers are better: IMD = -97.6dB; d2L = -108.6 dB; d2H = -102.5dB
LISTENING TEST
Some notes about listening. The unit plays currently in my room with the same set-up used for the listening of Threshold FET nine/e (tested here).
In the Stereophile’s review in the late ’80, we can read phrases like “the state-of-art-preamplifier” and “the most nearly perfect such device tested or read about elsewhere”.
Personally, if I had to use one single word to characterize it, I would use the term “transparent”, also according to the results of measurements above.
I was searching for any difference switching between two configurations (although not in a blind test): with and without the preamp in the chain i.e. RME ADI2 pro – FET Ten/e – Krell KMA160 and RME ADI2 pro – Krell KMA160. I was not be able to find anything of meaningful. The FET ten/e does not add or subtract nothing substantial to sound: resulting sound characteristics depend on other components in the chain.
So, I agree with Stereophile’s review: a high-class preamp that can be well used as reference to compare the behaviour of all others.
I tried to discover why, via measurements in different working conditions and listening tests.
The case of the FET Ten/e is well built, compact and robust. It has:
- Stepped Nobles balance and volume controls;
- Tape outputs (and related selectors);
- Selector switch for five inputs (no phono stage; the PC unit was sold separately);
- Attenuator of -20dB (no amplification) and Mono switch.
It has also one XLR input and both XLR and RCA outputs, presents only in the “e” version.
All FET Ten/e units are not equally; this unit should be the last version with military grade parts, only bettered by the extremely rare FET 10/t. Serial number is 064xx.
Its intern is amazing: all discrete components, separated board gain modules and an attention to all details. It uses discrete J-FET and MOSFET.
Power supply (15V) is made by an over dimensioned outboard (not shown).
In ’90, the price in Europe was about 3200€.
SPECIFICATIONS
Frequency response: 1.5Hz–150kHz +0, –3dB
THD: 0.01% at 1V out.
Max output: 20V
Gain: 20dB max
Output impedance: 100 ohms.
MEASUREMENTS
All following measurements have been made with:
- Input: XLR; from digital to analog with resolution 24/192
- Output: XLR; from analog to digital with resolution 24/192
- Load: 9 Kohm
FREQUENCY RESPONSE
Input: 1Vrms Tone sweep 10Hz-50KHz; Gain: 0dB (50% Volume) – LR Amplitude
The frequency response is excellent, flat far beyond the audible band, with a difference a little difference of 0.15dB between channels.
Input: 1Vrms Tone sweep 10Hz-50KHz; Gain: 0dB (50% Volume) – LR Phase
In the phase shift we see the same excellence: distortion is 0 also at 10Hz (wow), despite the low load, 9KOhm. This means a coherence for bass. Less than 3 degrees at 20KHz, very good.
CROSSTALK
Input: 0.1Vrms 10Hz-50KHz; Gain: 20dB (100% Volume) – LR/RL Crosstalk
Crosstalk is very good, even if there is an asymmetry of about 15dB among channels, maybe due to circuit layout. At 1KHz we have a very good -85dB (on L) and an optimum -99dB (on R).
Input: 1Vrms 10Hz-50KHz; Gain: 20dB (100% Volume) – LR/RL Crosstalk
Curiously, increasing the input of 20dB to reach 1Vrms (Gain still at 20dB, so output reaches 10Vrms) only the lower curve (L over R) shifts down, increasing the gap to 18dB.
Input: 1Vrms 10Hz-50KHz; Gain: 0dB (100% Volume); Attenuator ON – LR/RL Crosstalk
Switching ON the -20dB Attenuator, the curves became closer (5dB), but both shift up, to higher values.
SINGLE TONE DISTORTION
Input: 1Vrms @ 1KHz; Gain: 0dB (50% Volume) – LR Spectrum
At 1Vrms input, 0dB Gain, SNR is a good 83dB, with a TD+N of -86dB, for both channels. Harmonic distortion is mainly in the second harmonic (-100dB) and decreases with others (-111dB, -124db …), inaudible. Few the differences between the channels. Excellent the noise to power supply, practically absent, below 120dB.
Input: 0.1Vrms @1KHz; Gain: 20dB (100% Volume) – LR Spectrum
Maintaining fixed the output level at 1 Vrms and changing the Gain at 20dB and input level, SNR became 77.7dB, and TD+N of -79.6dB, for both channels. Harmonic distortion is due all to the second harmonic at -106dB, excellent. Unchanged the hum, always excellent.
Input: 1Vrms @ 1KHz; Attenuator: ON; Gain 0dB (Volume 100%) – LR Spectrum
With 1Vrms input and output but with -20dB Attenuator set to ON (Gain 0dB, Volume at 100%), the result is still very good: SNR is a good 83.2dB, with a TD+N of -86.7dB, for both channels. Harmonic distortion remains mainly in the second harmonic (-99dB) and decreases with others (-111dB, -124db …), inaudible. Few the differences between the channels; Unchanged the hum.
HARMONIC DISTORTION
Input: 1Vrms Sine @ 20Hz-20KHz; Gain: 0dB (Volume 50%) – THD per Frequency
At 1Vrms in the output, 0dB Gain (50% of volume) THD is 0.005% practically constant for all frequencies, below a very good 0.001% (-99dB), all due to second harmonic; the third one is about 0.0003% and it increases for values of frequency outside the audible band. The other harmonics are on lower values 0.00005%. THD+N is a good 0.005%. Excellent.
The graph (not shown) has not meaningful changes with 5Vrms input and Gain -14db (with or without attenuator).
Input: 0.1Vrms Sine @ 20Hz-20KHz; Gain: 20dB (Volume 100%) – THD per Frequency
In the standard condition, both THD and THD+N remain flat for all frequencies: about 0.0006% for THD (-104dB) and 0.01% (-80dB) for THD+N, excellent values. Distortion is mainly on the second harmonic; the others are practically inexistent.
Input: 0.12Vrms – 1.2Vrms Tone @ 997Hz; Gain: 0dB (Volume 50%) – THD per Level (R)
The distortion for level is also very good, with optimal values around 0.2Vrms output; above, increases practically only the second harmonic.
Input: 0.12mVrms – 0.12Vrms Tone @ 997Hz; Gain: 20dB (Volume 100%) – THD per Level (R)
Setting the Gain to max, and leaving the 1V as max output level, we have still very good results, with optimal values around 0.5Vrms output:
INTERMODULATION
Input: 0.7Vrms Sine @ 19 + 20KHz; Gain 0dB (Volume 50%) – Spectrum R
Very good also the CCIF test: IMD is -94.5 dB. Symmetrical nonlinearities (odd-order), near input signals, are about -104dB; asymmetrical, even-order distortion at 1KHz is -99dB; TD+N is -83dB. Negligible differences (not shown) between channels.
Input: 0.07mVrms Sine @ 19 + 20KHz; Gain 20dB (Volume 100%) – Spectrum R
With Gain at 20dB, 1 Vrms output, the numbers are better: IMD = -97.6dB; d2L = -108.6 dB; d2H = -102.5dB
LISTENING TEST
Some notes about listening. The unit plays currently in my room with the same set-up used for the listening of Threshold FET nine/e (tested here).
In the Stereophile’s review in the late ’80, we can read phrases like “the state-of-art-preamplifier” and “the most nearly perfect such device tested or read about elsewhere”.
Personally, if I had to use one single word to characterize it, I would use the term “transparent”, also according to the results of measurements above.
I was searching for any difference switching between two configurations (although not in a blind test): with and without the preamp in the chain i.e. RME ADI2 pro – FET Ten/e – Krell KMA160 and RME ADI2 pro – Krell KMA160. I was not be able to find anything of meaningful. The FET ten/e does not add or subtract nothing substantial to sound: resulting sound characteristics depend on other components in the chain.
So, I agree with Stereophile’s review: a high-class preamp that can be well used as reference to compare the behaviour of all others.
Last edited: