Amp Lab: 2DI4 Sonic Farm

This Article Was Originally Published On: July 1st, 2014 #Issue 14.

The Circuit

The signal path of the 2di4 is simple and clean. The 2di4 boasts a single ¼” input, which is buffered and folded back to an unbalanced ¼” jack labeled “AMP” on the front panel. This lets the musician connect to an amplifier with a buffered, instrument-level signal. This is a really nice touch, and a welcome part of this DI.

The 2di4 offers four independent means to shape the gain structure through the DI circuitry. The four means include a trim switch, a GAIN button, a Triode/Pentode (T/P) button, and an Output knob. The 2di4 offers a first means controlled by a pad switch that allows the user to select one of three options: 0dB, -12dB or Mute. Flip the pad switch to the mute position for quick instrument changes, use the -12dB position to accommodate low-headroom devices downstream of the 2di4, or leave the pad switch in the 0dB setting to effectively bypass attenuation at the pad switch circuitry. Switching between Triode and Pentode mode (discussed below) can sometimes cause a pop because of the way that the circuitry is reconfigured. As such, switch the pad switch to the mute position when toggling the T/P switch.

A second means is controlled by a GAIN button on the front panel. Depressing the GAIN button affects the signal level, adding a fixed gain. The 2di4 features a single active gain stage, which is implemented by an EF 86 vacuum tube. The gain added by the GAIN button is created by bypassing the current setting circuitry designed around the cathode of the EF86.

A third means is also built around an EF86 vacuum tube. The EF86 is normally a pentode vacuum tube that can be operated in either triode mode or pentode mode, selected based upon a Triode/Pentode (T/P) switch. In short, the T/P switch “rewires” the tube to operate in pentode mode or triode mode. The T/P switch also reconfigures the bias for proper operation. The triode and pentode modes are discussed in greater detail below.

A fourth means is the Output control, which is adjusted by the output potentiometer. The EF 86 has a high output impedance. As such, the tube output is buffered before the signal is coupled to the XLR output jacks on the back panel. The Output control is after the gain is applied, so tube gain and tube overload are unaffected by the Output control. However, the Output control does affect the level at which the output buffer, transformer and connected output device are driven.

The 2di4 also offers a few equalization options, including two low-frequency options, selectable with a LO switch, and two high-frequency options, selectable with a HI switch. A neat trick here is that the tonal shaping is performed by providing frequency dependent elements (capacitors for the HI boost, and inductors for the LO boost) in the cathode circuitry of the EF86 tube. The benefit here is that the circuitry is clean and the signal path is as short as possible. The tone-shaping elements are not in the direct signal path. The disadvantage here is that because the GAIN, LO and HI all operate on the cathode circuitry around the EF86 tube, the GAIN switch, when engaged, defeats the LO and HI tone-shaping options. As such, you have the option of no gain/no EQ; gain with no EQ; or  EQ.

Referring to Fig. 2, a frequency response plot shows that the LO and HI options each provide subtle shelving boost relative to the flat setting (LO and HI switches in the center position). Although subtle, the EQ is effective in its role as a DI tone control. If you need major equalization, you need to make adjustments at the source (or replace the source). Note that the response illustrated in Fig. 2 is merely illustrative of the shape of the EQ curves. The ultimate level of boost is determined by a number of factors, including other DI settings. In this regard, the controls are interactive and work together. Moreover, there are two small trim controls (labeled TRIM – located between the T/P switch and Output knob) that you can tweak with a small screwdriver to adjust the overall level, to obtain slightly more, or less, gain.

FIG.2 EQ Options

The output of the EF86 tube gain stage is buffered, and the buffered output is coupled to a Cinemag transformer-balanced, DI output that connects to a first XLR jack on the back panel. The buffered output of the EF86 also feeds an op-amp balanced line output that is connected to a second XLR jack on the back panel. As such, the 2di4 provides an instrument level, a transformer-balanced microphone output and a line output, which are simultaneously available.

A signal indicator designated OL is located between the input jack and the GAIN button. The LED glows Green if a signal is detected as being present. If the signal exceeds a predetermined threshold, then the LED glows Red.

To get a sense of the gain of this device at both the line and mic levels, we set the trim switch to its 0dB position, and set the GAIN and T/P switches out. We calibrated the line level out to provide 0dB as our baseline and ran a frequency sweep. Regardless of switch settings, the frequency response was similar. We measured approximately 28dB difference between the mic and line levels throughout the tests. Under our test conditions, engaging GAIN switch resulted in about 4dB gain. Engaging the T/P switch resulted in about 7dB gain. However, engaging both the GAIN and the T/P switches resulted in just under 18dB gain.

Switching our tests, we measured about 35dB gain in pentode mode and approximately 28dB in triode mode, with a THD+N under 1%. The nature of the particular design in this product allows for measureable distortion levels well over 1% without any sign of nasty distortion. I would not hesitate to call the signal “clean” at levels over 2% THD+N. In this regard, gain in pentode mode can reach into the 50+dB range.

Triode/Pentode Switch

Before getting into the function of this switch, let’s take a minute and address this whole pentode/triode thing. Let’s start with a triode, because that will make the pentode easier to understand. A triode has basically three controllable electrodes (hence tri-ode), including an anode, cathode and grid. Each electrode is accessed by the designer as a pin on the tube. A triode also has heaters that heat the cathode, but let’s not concern ourselves with them, here, for this discussion. Basically, the heated cathode releases electrons that are attracted across the vacuum of the tube to the anode (an electron migration). An input voltage applied across the grid is injected into this electron migration, such that a small change in input voltage is converted into a large change in current. Circuitry attached to the anode converts the change in current to a voltage.

Now, a pentode is basically a triode with two additional electrodes (which we see as two additional pins on the tube), including a screen grid and a suppressor grid. Without getting too techy, the screen grid provides two functions. It accelerates electrons towards the anode, thus providing increased gain over a conventional triode. Moreover, it reduces capacitance that forms between the grid and anode. However, the screen grid adds artifacts that are undesirable. These artifacts are corrected for by the suppressor grid. Now, the neat trick is that if the screen grid is connected to the anode, and the suppressor grid is tied to the cathode, the resulting configuration is a triode.

Life does not come without tradeoffs and compromises. Even though the Pentode position gives more gain, you have to give something in return, and that something is headroom. To see these tradeoffs, take a look at Figs. 3-5. Under our test conditions, we set the trim switch to 0dB, we set the GAIN switch to its out position, we dialed the Output potentiometer to noon and swept our input signal level from off to 5 Vrms. Pentode mode is illustrated in the red trace and triode mode is illustrated in the blue trace.

As illustrated in Fig. 3, for low signal levels, the Pentode mode has about 7dB more gain compared to the Triode mode. However, the gain of the Pentode mode starts dropping off at about 1.5 Vrms (indicating the onset of clipping), whereas the gain in Triode mode stays clean up to an input signal of about 3.5 Vrms.

FIG.3 Gain

As illustrated in Fig. 4, for low signal levels, the Pentode mode and the Triode mode are extremely linear, tracking gain as a function of input signal with precision. Fig. 4 is consistent with Fig. 3 in that the difference in gain between Pentode mode and Triode mode is about 7dB, with Pentode mode providing more overall gain. However, the Triode mode demonstrates consistently more headroom. Note that when reading a linearity graph, distortion begins to onset when the slope of the line begins to flatten out.

FIG.4 Linearity

As illustrated in Fig. 5, the tube characteristics of this device are clearly apparent. The design of the circuitry in the 2di4 allows the THD+N measurements to reach a few percent before you would even begin to notice. In our tests, THD+N as high as 2-3% yielded no signs of clear, nasty clipping. For low signal levels, both Pentode mode and Triode mode are clean at well under 1% THD+N. However, in Pentode mode, under our test conditions, the distortion quickly increased for inputs above about 1.5 Vrms, whereas the signal remains relatively clean in Triode mode to over 3 Vrms. To illustrate this, refer to Fig. 6. We input a 3.15 Vrms input signal in Triode mode and measured about 2% THD +N. As the scope trace illustrates, whereas a critical eye might suggest that there are signs of compression rounding out/fattening up the peaks, there are no signs of hard clipping.

FIG.5 THD+N ratio

FIG.6 Scope Trace Triode Mode 3.15 Vrms Input

Keep in mind that headroom is not the only difference between the Pentode mode and the Triode mode. Each offers its own harmonic signature. For instance, compare Fig. 7 to Fig. 8. Fig. 7 shows the 1% THD+N in Triode mode, whereas Fig. 8 shows 1% THD+N in Pentode mode. Both modes show that the second harmonic is about 40dB down. However, the Triode mode harmonics show a gradual “stair step” response, with the third harmonic over 30dB down, compared to the second harmonic. Comparatively, in Pentode mode, the third harmonic is about 15dB down from the second harmonic. Moreover, the higher-order harmonics exhibit a relatively flatter response.

FIG.7 1% THD+N triode mode

FIG.8 1% THD+N pentode mode

Due to gain differences, biasing differences and other effects of the tube configuration, the Triode/Pentode switch offers not only the opportunity to trade off input gain for headroom, but also to alter the harmonic signature of the device.

Conclusions

On the bench, the 2di4 appears to be a strong contender for supreme DI. Yes, I am impressed with this DI. Is it 2 di 4? Well, as far as DI boxes go, this box should deliver the goods. If you own a studio, this DI should be on the short list of additions (assuming that you do not already own  it). If you play bass, the wide range of headroom, the tube character, the subtle EQ shaping, the instrument out, DI out and line level out make this product extremely versatile.

Do you truly need a DI will all of these bells and whistles? Maybe not. But if you are looking for the ultimate DI, I would start, and likely end, my search right here.

If I had one complaint, it would be that the overload indicator was not as precise as I had expected. It seemed to sometimes “latch” onto an overload indication, even when the input signal clearly fell below clipping. Also, under certain conditions, the overload indicator turns red when the observable scope trace looks clean. As such, I would use the overload indicator as a “rough approximation,” but I would trust my ears over the LED color when determining whether an adjustment is necessary. Then again, I think that is good practice, no matter what the gear.

1-5 (unacceptable to impeccable)

On-bench

SONIC PROFILE:

Low: full, tight, controlled
Mids: clean, pure, slightly warm
Highs: smooth, detailed, inviting