Threshold FET Nine/e (vintage) Preamplifier Review

[Pinox67]

I share in this post the review of a vintage preamplifier, the Threshold FET Nine/e. A project by the genius of Nelson Pass, thirty years old; price was about $2500.Thus, this is not a review for a “news”; however data below could be useful to make comparisons with modern preamplifiers.

This unit belongs to a friend of mine that asked to me for a check. In fact, it has been “refurbished” by a technician for the volume knob. I’ve done both measurements and a listening test described below.

The chassis is compact and very robust, with an outboard power supply (not shown in the picture). Its amplification is managed by J-FET and MOSFET. It has also a phone stage; input/output connectors are only of RCA type.

SPECIFICATIONS (for line stage)
Frequency response: DC–100kHz, –0.2dB
Distortion: 0.007%;
Input impedance: 20 kohm;
Noise: –112dB ref. maximum output.
Output impedance: 100 ohm.

MEASUREMENTS [EDITED May 25, 2020]
All following measurements have been made with:
- input: digital to analog (RCA), with resolution 24/192
- output: analog (RCA) to digital, with resolution 24/192
- load: 9 Kohm

FREQUENCY RESPONSE

Input: 1Vrms Tone sweep 10Hz-96KHz; Gain: 0dB (50%) - Amplitude

Input: 1Vrms Tone sweep 10Hz-96KHz; Gain: 0dB (50%) - Phase

The frequency response is very flat, with a difference of less 0.1dB between channels: Very good!
The phase response is near perfect too: it’s no common to find such behaviour. This means coherence of sound in all spectrum, no delays. The behavior is also confirmed checking the output with a square wave, near perfection:

Input: 0.7Vrms Square Wave at 50Hz; Gain 0dB (50%)

Do not consider the spikes due to the digital conversion.

CROSSTALK

Input: 1Vrms 10Hz-96KHz; Gain: 20dB (max); CC

Perfectly symmetric the effect in both channels. We have at 1KHz a good 82dB, but at higher frequencies it rises quickly, due to capacitive coupling between channels. But the values are still good.

TONE DISTORTION

Input: 1Vrms at 1KHz; Gain: 0dB (50%)

Input: 1Vrms; Gain: 20dB (100%)

SNR is about 84dB and TD+N 0.004%, overall not so bad, distortion+noise is not audible (at least at 1KHz).
Distortion is mainly due to second and third harmonics; when gain is increased at 100% (20dB), increases also higher lever harmonics, not monotonic with frequency. Values are however still low, so no harsh of sound is audible.
At 100% there are evident some harmonics due to AC (at 50Hz), still low, but could become audible is some set up.

HARMONIC DISTORTION

Input: 1Vrms 20Hz-20KHz; Gain: 0dB (50%)

The overall Distortion+Noise, THD+N, is good: about 0.004% at 1KHz. But it increases of 10 ten times at 20KHz. Distortion is mainly due to second and third harmonics on all frequencies, even if it increases more at high frequencies. Harmonic distortion, THD, is 0.001%, at 0dB. So, noise is (very) predominant.

Input: 0Vrms – 3Vrms Tone at 997Hz; Gain:0dB (50%)

The total distortion, THD+N, is good at 1V, about 0.006%; it raises to 0.07% at 100mV and to 0.7% at 10mV. Up to 200mV, harmonic distortion is hidden from noise; THD is about 0.0005%; above 200mV it raises, due to mainly to second harmonic. This up to 2.3V, the maximum level that I measured.

TRANSIENT

Input: 1Vrms burst of 6.5 sine at 1KHz; Gain 20dB (max)

All cycles have same amplitude and there is no trace of memory or oscillation: Perfect!


INTERMODULATION: CCIF

Input: 0.7Vrms Sine 19+20KHz; Gain 0dB (50%)​

Very Good at 0dB Gain: IMD is 0.0018%, with very low intermodulation components, not audible.


SUBJECTIVE SOUND QUALITY [EDITED May 25, 2020]
The unit has played in my (dedicated) room for some days. The set-up is composed of: as digital source, RME ADI 2 Pro, connected to a pair of power amplifiers Krell KMA160; Wilson Audio Sasha II as loudspeakers. AudioQuest cables. Music genre: rock, pop, r&b, jazz.
Below the listening impressions by more persons. The FET nine has been compared with the two good tube based preamplifiers and one passive preamplifier, listening to same (few) tracks. However, the test is not blinded and level matching is by ear.

What I can say is that the FET nine, in the described set-up, sounds very, very good: it is clean and flowing; I can detect a bit of noise floor in the loudspeakers, but it is very low. I would say that:

- Soundstage is excellent, a lot over the average: sound objects can be well distinguished in the space in all directions.
- Bass are well present, controlled and coherent.
- Medium have a good presence and they are detailed.
- High are airy and smooth, not so far from tube sound.
- The sense of realism is good, even if it is desirable a little more dynamic contrast to have the illusion of a real live event.

As final words, I would say that it is a fine and elegant device. It has a engaging sound, and mostly not tiring in long listening sessions. It compete without any doubt with a lot of modern preamplifiers with a price of several thousand of dollars that I heard.

Any question or comment is welcome.

 

[Apesbrain]

I love mine! Same color. (It was available in black and gun metal as well.) Also have the companion SA/3.9e power amp (60 class A WPC). Thanks for the review.

Here's a picture of the pretty internals:

 

[SIY]

​

The frequency response is very flat, with a difference of less 0.1dB between channels: Very good!
The phase response is near perfect too: it’s no common to find such behavior! This means coherence of sound in all spectrum, no delays.

If frequency response is flat, so is phase response unless someone deliberately builds in a delay line. Nothing remarkable.

​

SNR is about 84dB and TD+N 0.004%, overall not so bad (SINAD 88dB), distortion+noise is not audible (at least at 1KHz).
Distortion is mainly due to second and third harmonics; but it is not monotonic with higher harmonic when gain is increased up the maximum, 20dB (measured). This can make the sound harsh. At those levels? That's a highly unlikely claim.
At 100% also a lot of harmonics due to AC (at 50Hz) are evident and could become audible. At those levels? That's a highly unlikely claim.

​

The total distortion, THD+N, is good at 1V, about 0.006%, but it raises quickly at lower levels: at 100mV is 0.07% (audible) and 0.7% at 10mV. So, we don’t expect a unit able to reveal fine details at low volume. Another highly unlikely claim.


SUBJECTIVE SOUND QUALITY You don't mean "subjective," you mean "done without using even a minimum of basic controls."
The unit has played in my (dedicated) room for some days, tested with a digital source (RME ADI 2 DAC) and connected to a pair of Krell KMA160; Wilson Audio Sasha II as loudspeakers. AudioQuest cables. Music genre: rock, pop, r&b, jazz.
After a deep listening sessions, I can say that it sounds very, very good: it is clean and flowing; I can detect a bit of noise floor in the loudspeakers, but it is very low. I would say that:

- Soundstage is excellent, a lot over the average. That is unbelievable, in a literal sense. Sound is fully 3d; sound objects can be well distinguished in the space in all directions.
- Bass are well present, controlled and coherent.
- Medium have a good presence and detailed.
- High are airy and smooth, not so far from tube sound.
- The sense of realism is good, even if it is desirable a little more dynamic contrast to have the illusion of a real live event. Zero to do with the preamp unless a preamp design is badly broken.

As final words, I would say that it is a fine and elegant device. It has a engaging sound, not tiring also in long listening sessions. If it has a "sound," it's badly broken. Your measurements are evidence that it's not and that this sort of "observation" is a result of the total lack of basic controls in the listening evaluation.It sounds far better than a lot of modern preamplifiers with a price of several thousand of dollars that I heard. Extraordinary claim given with zero evidence.

Any question or comment is welcome.

Comments in red. I hope you really meant your last sentence.

 

[Pinox67]

Hi SIY,
Thanks for your note. Some clarification to explain the misunderstandings.

About the Phase response, I facing sometime not a flat response, even if Amplitude is flat. For example, this is a measurement of Jeff Rowland Consonance:

Amplitude is flat (red line), but phase (green line) is not. Artefacts of measure via sweep?

About distortions: maybe I said that badly about harsh: this is not the case; it is a general statement. Increasing volume levels (as far as possible) no harshness is perceived; only a minor noise on the loudspeakers.

About fine details: This preamp need a good volume level to shine; at low volume is less engaging respect to other preamps. So, I suspect that this could be due to this measure, not excellent.

About the "basic control" you mention for Subjective Sound Quality: I don't understand well what you mean.
What is described here is simply the sensations of auditors (me and some friends, with different tastes and training) listening to sound resulting from the described set-up (source, power amp, loudspeaker, cables and room).
And more auditors listened to more set-up in the same room changing only the preamp. We'd the opportunity to test preamps like AudioNote M8, Conrad Johnson PV10, Jeff Rowland Consonance, Mark Levison n.32, NAD S100, British Fidelity MVT, Angstrom Lab Stella, Audio Reseach LS2 and LS25, Xindak XA3200 and others.
Each one results in a different sound with the described set-up and it is shining is some aspects.
The Threshold shines in the soundstage, while other preamps are less strong in this aspect. Probably with a different set-up the result could be different. Here is stated what we experienced. Very simple!

 

[SIY]

Hi SIY,
Thanks for your note. Some clarification to explain the misunderstandings.

About the Phase response, I facing sometime not a flat response, even if Amplitude is flat. For example, this is a measurement of Jeff Rowland Consonance:

Amplitude is flat (red line), but phase (green line) is not. Artefacts of measure via sweep? Possible. Or (less likely) they put a sharp ultrasonic filter above the high end of your measurement. I'd examine the measurement a bit more closely (phase should go as the first derivative of the magnitude).

About distortions: maybe I said that badly about harsh: this is not the case; it is a general statement. Increasing volume levels (as far as possible) no harshness is perceived; only a minor noise on the loudspeakers.

About fine details: This preamp need a good volume level to shine; at low volume is less engaging. Thus, this could be due to this measure, not excellent. The measurement looks fine. This preamp is almost surely acoustically transparent.

About the "basic control" you mention for Subjective Sound Quality: I don't understand well what you mean. Any by-ear subjective evaluation has to be double blind and level-matched (ears only, no peeking). otherwise, it's totally unreliable.
What is described here is simply the sensations of auditors (me and some friends, with different tastes and training) listening to sound resulting from the described set-up (source, power amp, loudspeaker, cables and room).
And more auditors listened to more set-up in the same room changing only the preamp. We'd the opportunity to test preamps like AudioNote M8, Conrad Johnson PV10, Jeff Rowland Consonance, Mark Levison n.32, NAD S100, British Fidelity MVT, Angstrom Lab Stella, Audio Reseach LS2 and LS25, Xindak XA3200 and others.
Each one results in a different sound with the described set-up and it is shining is some aspects.
The Threshold shines in the soundstage, while other preamps are less strong in this aspect. Probably with a different set-up the result could be different. Here is stated our experience. Very simple! You confirmed that the basic controls weren't there. This was NOT ears-only and level matched. I would strongly suggest you try comparing a couple of them with controls. You'll have a revelatory experience, I promise you.

Comments in red again. And sincere thanks for taking the criticisms in the intended spirit; I'm always delighted when people actually do measurements, and I'd like to persuade you to try a well-controlled listening test.

 

[Willem]

This is of course a nice piece of gear. Even so using the ADI as a pre amp would be a lot simpler.

 

[GXAlan]

What are you using to make your measurements?

 

[Pinox67]

Comments in red again. And sincere thanks for taking the criticisms in the intended spirit; I'm always delighted when people actually do measurements, and I'd like to persuade you to try a well-controlled listening test.

Criticisms are welcome if they are motivated and finalised to the (mutual) grow.

For the measures above, I think that overall they are very good, no flaw.

For "basic control" in the listening tests, I understand what you mean.
Well, tests have been executed with a close comparison (1-2 minutes for the switch) with a reference preamp, listening to the same (few) audio tracks of different types. The test is not blind. Level matching is approximative, by ear.
We'll try to execute it as you suggest, with more parameters controlled.

What do you use to calibrate quickly the exact level for the tests?
Do you measure in advance the volt on output with a test signal?

 

[Pinox67]

What are you using to make your measurements?

I use REW SW + RME ADI2 Pro FS (calibrated) and a multimeter.

 

[Pinox67]

This is of course a nice piece of gear. Even so using the ADI as a pre amp would be a lot simpler.

Yes, I tried also this set-up and the results are very good.

 

[SIY]

Criticisms are welcome if they are motivated and finalised to the (mutual) grow.

For the measures above, I think that overall they are very good, no flaw.

For "basic control" in the listening tests, I understand what you mean.
Well, tests have been executed with a close comparison (1-2 minutes for the switch) with a reference preamp, listening to the same (few) audio tracks of different types. The test is not blind. Level matching is approximative, by ear.
We'll try to execute it as you suggest, with more parameters controlled.

What do you use to calibrate quickly the exact level for the tests?
Do you measure in advance the volt on output with a test signal?

I would start by doing a bypass test. Use a voltmeter to carefully set the gain to 1. You can do that by connecting your RME to the preamp input, generating a sine wave at (say) 500Hz, measuring the input and output levels, and setting the preamp's volume control to get an equal output. You can then set up the signal flow through the tape monitor so that the active circuitry can be bypassed or switched in. Confirm with the voltmeter that the signals bypassed and through the active circuit are still equal. Switch it back and forth a few times. You may find right then and there that the circuit is inaudible to you.

If you think you still hear a difference, you can then set up something to obscure your view and have someone out of your sight (leave the room if necessary) set the switch to one position or the other. Have that switch position randomly chosen (your assistant can do coin flips each time for good randomizing and keep a key sheet) and try to identify "in" or "out" 12 times. Keep track of your guesses, then after the experiment is over, compare your guesses to the key sheet.

An alternative way (and the way I'd do it), since you have a superb audio interface, is to do the level match, record whatever music or other selections you like with the unity gain preamp in the circuit and again with the unity gain preamp out of the circuit. Then use Foobar or a similar ABX software to see if you can hear that difference when peeking is impossible.