Amp Lab: Acoustic Image Flex Preamp

This Article Was Originally Published On: February 3rd, 2015 #Issue 15.

Construction

The front and rear of the AI Flex preamp housing is made from aluminum, but the bottom is made from steel (to facilitate magnetic attachment to the Flex Cab). The overall form factor strikes a hybrid appearance between a mixing console and an instrument amplifier. When sitting on a flat surface, the AI user controls sit at an angle like a mixer. However, the two channels run horizontally across the housing, as opposed to vertically, as you would find on many analog mixing consoles. The output jacks are provided on the back panel in a patch bay, much like you might find on a conventional mixing console.

Referring to Fig. 1, we provide a photo illustration of some of the user-accessible jumpers that can be used to configure hidden features of the Flex preamp. The screen printing is pretty small, so we have zoomed in on one region of the circuit board to show the labeling. The jumpers will be discussed in greater detail, below.

Fig 1. Jumpers

The Tone Stack

The Flex includes two input channels that eek uber-clean vibes. Our test system measured 20Hz – 20kHz +/- 0.6 dB; 200 mV swept sin input, with a limited bandwidth measurement of 80Hz – 8kHz +/- 0.07 dB; 200 mV swept sin input. If you want clean and flat, set all of the EQ controls to their center detent (noon) positions, bypass the variable high-pass filter, and jam on. No need for an “all controls at noon” vs. “optimally flat” plot, here.

Referring to Figs. 2-5, we provide EQ plots for each of the included EQ bands. In Fig.2, the Low band shows a clear shelving function, providing over 12dB of boost and cut at about 50Hz. In Fig. 3, the Lo Mid shows a bandpass function, providing just over 12dB of boost or cut at just under 300Hz. In Fig. 4, the Hi Mid shows a bandpass function providing just over 12dB of boost or cut at just over 1.2kHz. In Fig. 5, the High band shows a clear shelving function, providing in excess of 12dB of boost and cut at about 8kHz.

Fig 2. Low sweep

Fig 3. Low Mid sweep

Fig 4. Hi Mid sweep

Fig. 5 High sweep

Referring to Fig.6, we matched the gains as best we could between Channel 1 and Channel 2. We then ran the EQ of Channel 1, compared to Channel 2. As can be seen from the plot (which only shows the Lo Mid and Hi Mid bands), the two EQ sections are, for all intents and purposes, identical. Kudos, here, for maintaining such tight control over the tolerances during manufacturing.

Fig 6. EQ1 compared to EQ2

Referring to Fig. 7, we show a plot of the variable high-pass filter engaged. As noted in the plot, there is a slight drop in overall level when the filter is engaged. Here, the black trace is with the variable high-pass filter bypassed. The black trace is a couple of dB higher in gain than the remaining traces, which show various filter settings with the variable high-pass filter engaged.

Fig 7. Variable high pass filter

There is little range of adjustment from its fully counterclockwise (Off) position (red trace) to the 9:00 setting (blue trace). However, from the 9:00 setting to fully clockwise, the plot demonstrates that the variable high-pass filter, extending from about 30Hz (blue trace) to just over 150Hz (green trace). Note, here, that there appears to be a wide range of adjustment between the noon setting (dark blue trace) and 3:00 setting (mustard yellow trace), so you may want to really work this range for dialing in a tight tone.

I have said this before, and I will say it again. The variable high-pass filter control is, in my opinion, one of the most powerful and useful controls that a preamplifier can have. It seems to be the Gram Parsons of EQ controls … under-rated by those that do not understand it.  

Input/Output

There are a number of ways to connect the Flex Pre outside of its use with the Flex Cab. The easiest is to use the mono Preamp out on the patch bay. The Preamp out is a ¼” jack that connects to a “mixbus” inside the Flex preamp. The mixbus is the internal circuitry that mixes Channel 1 and Channel 2. Another way to access the output is through a Signal/Power Interface output. This is a proprietary output on a CAT-5 connector (typically used to connect computers in the caveman days of wired networks). Fortunately, AI provides a CAT-5 to ¼” jack converter (as well as a CAT-5 to CAT-5 cable). However, as we will discover, all is not created equal in the patch bay.

Input Gain

Let’s start off on the input side of things. The Flex includes two channels, each with a single XLR and a single 1/4” jack. There is a 10dB gain switch for the XLR connector. However, there is no pad switch for the ¼” input. Also missing is the familiar option for selecting between active and passive inputs. Rather, the 1/4″ input provides a relatively high impedance in excess of 1 MOhm. Accordingly, we wanted to explore the range of inputs that the FLEX preamp can handle.

With the Level control at about 9:00, the output is just under unity, give or take a bit. Pushing the input, we started observing clipping at about 1.2 Vrms, 400Hz sin input. When using the Preamp out jack on the patch bay, the clipping appears first on the negative-going transitions. As the signal level at the input continues to increase, hard clipping appears on both the positive and negative-going peaks. Fig. 8 illustrates the onset of hard clipping on the negative-going peaks.

Fig 8. Onset of clipping

Correspondingly, when using Signal/Power Interface output with the provided conversion cable, under the same test conditions, we achieve a similar ceiling of about 1.2 Vrms, 400Hz sin input. However, this time, the clipping begins to occur on the positive-going edge, as illustrated in Fig. 9. As the signal level at the input continues to increase, hard clipping appears on both the positive and negative going peaks.

Fig 9. Onset of clipping adapter out

The Level control appears to show little to no difference in signal handling capability between off (fully counterclockwise) and about 9:00 or just under. That is, we could not improve upon our clean headroom significantly, by turning the Level control down lower than 9:00. Indeed, no signal will pass when the Level control is fully counterclockwise (Off).

As such, from our bench tests, it appears that if your bass uses active pickups with strong output, you run a risk of overloading the preamp on the peak transients, given the basic configuration. Signal headroom can be improved using the limiter and/or a jumpered pad, each of which is discussed below. 

Output

With the input Level control at its maximum value, we were able to achieve clean gain of approximately 5.5 Vrms output with a 1 Vrms, 400Hz, sin input at 1.7% THD+N (clean looking on the scope), using the Preamp out jack. In this regard, pay attention to the level when you are interfacing this with amplifiers that are not part of the Flex family, as there is a lot of gain to be found on tap. This is particularly important when using the Preamp out jack as the limiter is unavailable on this output.

When using the Signal/Power Interface output, with the input Level control at its maximum value, we were able to obtain the same clean gain of approximately 5.5Vrms output. However, the 5.5 Vrms output was obtained using a 700 mVrms (compared to 1 Vrms using the Preamp out jack), 400Hz, sin input at 1.7% THD+N (clean looking on the scope).

Interestingly, we note that the Master volume control does not affect the clean headroom of the Flex preamp when using the Signal/Power Interface output. Rather, the Master appears to merely scale the signal. Thus, if the signal clips due to the input signal relative to the Level control, the Master will make that same clipped signal either larger or smaller in value.

Referring to Fig.10, we see a key difference between the Preamp output and the Signal/Power Interface output (no limiter). The plot of Fig.10 is a stepped level sweep. Basically, we set the input at a low value and capture the output. We then step up the signal level input to the Flex preamp and capture the output. This process is repeated over a range of input levels. The nice thing about this tool is that you can see where the device under test begins to clip or otherwise run out of clean headroom by noting the “elbow” where the gain begins to drop off.

Fig 10. Gain comparison

In all traces, the Level control was at max (fully clockwise). The black trace shows a level sweep from the Preamp output jack. This is our baseline, which shows a gain of about 15dB for all signals below about 1.2 Vrms. Input signals above this value begin to clip, causing the Flex to generate less gain.

The red trace, blue trace and green trace show the same sweep, this time using the Signal/Power Interface output. The red trace is the Master fully clockwise (full on). The blue trace is the Master at 3:00 and the green trace is the Master at noon. Note that using the Master, the gain (from the settings in this test) ranges from just over 10dB to just over 18dB up to about 800 mVrms. After about 800 mVrms, the gain starts to decrease. Note that using the Signal/Power Interface output causes the “elbow” (transition into clipping) to occur at approximately 800 mVrms, regardless of the Master setting.

Limiter

Use of the limiter is an effective way to deal with large input signal levels. To see the effectiveness of the limiter, we set the Level control to noon and Master at full on. Here, the Master control setting does effect limiter performance (but not headroom). We turn the Master full up to really exaggerate the limiter’s performance for purposes of illustration.

We set the input to 400Hz, and swept the input signal level. Fig. 11 shows a 200 mVrms 400Hz signal. At this level, it is clear that the limiter is kicking in by the observation that the output has a THD+N of over 5%. Without the limiter, the signal remains clean, under 2% THD+N, and typically under 1% THD+N. However, the signal appears, at least visually, to “look” like a sine wave. Turning to Fig. 12, at 315 mVrms, there are clear signs of the limiter rounding out the positive and negative-going peaks. Referring now to Fig. 13, by 500 mVrms, the THD+N level has risen to over 21%. Despite such high distortion levels, the signal appears to continue to be rounded off, with no sign of hard clipping. Referring to Fig. 14, pushing the input level to 1.5 Vrms (a reasonable peak transient for many passive pickups), we see the clear signs of hard clipping.

Fig 11. Scope monitor 400Hz 200 mVrms 5.3 THD+N

Fig 12. Scope monitor 400Hz 315 mVrms 13.82 THD+N

Fig 13. Scope monitor 400Hz 500 mVrms 21.2 THD+N

Fig 14. Scope monitor 400Hz 1500 mVrms 31.6 THD+N

Turning down the Level control and/or the Master control does not affect the overall headroom of the Flex. Even with the Level control and Master control below 9:00, a 1.2 Vrms input signal began showing trace evidence of hard clipping with the limiter on. Clear hard clipping was evident by 1.5Vrms.

It is also possible to push a relatively low input signal into hard clipping using the gain on tap with the Level control. For instance, up to about 900 mVrms, the Signal/Power Interface output (limiter engaged) will not hard clip, even with the Level control and Master control at Max.

However, at 1 Vrms input, with the Level control at noon, the Master control is effective at transitioning the output of the Signal/Power Interface output (limiter engaged) from clean (264 mVrms output at 1.7% THD+N) clear up to significant limiting (1 Vrms output at over 20% THD+N) with no sign of hard clipping. Comparatively, turning the Level control up to full on causes hard clipping with a 1 Vrms input signal, regardless of the Master control setting.

Jumpers

On the underside of the circuit board, there are five user-settable jumpers. A group of these jumpers is illustrated in Fig. 1. A first jumper provides a pad to the ¼” input of Channel 1. Likewise, a second jumper provides a pad to the ¼” input of Channel 2. To test the pad feature, we set the Level control at noon and the Master control at noon. We disengaged the limiter and swept a 400Hz sin from 10 mVrms to 1.6 mVrms. As illustrated in the plot of Fig. 15, there is approximately 5dB of padding with the jumper set to the Pad position. Here, the black trace is with the jumper in the normal setting, and the red trace is with the jumper shifted over to the pad setting.

Fig 15. Gain Normal vs Pad

The Pad setting has an interesting effect on the output of the unit. We set the Level control to 9:00 and the Master control to Noon. The limiter was turned off. Using the Signal/Power Interface output (limiter off), we applied a 1.25 Vrms signal, which previously clipped in our tests. The output appears clean as illustrated in Fig. 16. Referring to Fig. 17, cranking the input clear up to 2.5 Vrms, the lower peak shows clear rounding without hard clipping (this is with the limiter off). This shows clear asymmetrical distortion setting in. Cranking the input to a whopping 3.15 Vrms, the positive going peak show signs of hard clipping, as illustrated in Fig. 18. However, the negative-going peak continues to round off, showing signs of limiting rather than hard clipping. Switching to the Preamp output, this observation is flipped. The positive going peak rounds off limiting, whereas the negative going peak hard clips at 3.15 Vrms.

Fig 16. Scope monitor 400Hz Pad engaged 1.25 Vrms

Fig 17. Scope monitor 400Hz Pad engaged 2.5 Vrms

Fig 18. Scope monitor 400Hz Pad engaged 3.15 Vrms

A third jumper is provided to convert the Phase switch into a speaker emulator button on Channel 1. A corresponding fourth jumper is provided to convert the phase switch of Channel 2 to a speaker emulated output. Referring to Fig. 19, when the Phase switch is disengaged (red trace), the frequency response is normal, full range. When the phase switch is engaged (black trace), the output drops about 4dB for low frequencies. There is a small resonant peak around 3kHz with a treble rolloff above this.

Fig 19. Speaker Emulator

This is a really cool feature and a really useful way to get some clever tone settings. Combining the speaker emulator with the variable high-pass filter, a tremendous number of cool possibilities exist in a single knob solution. I do think it is a bit confusing to remember that “phase” = speaker emulator. In this regard, if the panel were labeled more clearly (e.g., “Phase/Emulator”) this feature may be more clear to users. That said, we realize that the front panel is already crowded and adding additional labeling may clutter things up too much.

The fifth jumper removes Channel 2 from the main mixbus and routes the output directly to a dedicated Preamp output on the patch bay. This is a cool way to dedicate Channel 1 to one Preamp output jack and Channel 2 to a second Preamp output jack, with no mixing within the Flex.

Conclusion

The Flex Pre was very stable and consistent in bench testing. For many users, the Flex Pre will provide the right amount of necessary features. The addition of a two-channel mixer, XLR microphone inputs, and built-in effects show that this unit is way more capable than the garden-variety bass amp that throws in the token iPod 1/8” input. One interesting feature worth mention is that the Flex preamp also comes with an adapter, allowing the unit to be mounted on a music stand or microphone stand for positioning close to the performer. This shows that AI intends the Flex to be a “hands-on” product.

 My biggest gripe with the Flex Pre is that the Signal/Power Interface does not mirror the Preamp output jacks. For instance, the polarity appears to be reversed between the two outputs, as suggested by Figs. 8 and 9. Moreover, the limiter does not function when using the Preamp out jacks. If you want the limiter to function, you must use the adapter. This is not spelled out clearly in the manual, and minimizes the functionality of the Preamp out jacks.

Another minor complaint is that the unit provides three feature options via jumpers accessible on the underside of the circuit board. I get that most users will “set and forget” the settings, and many users will not likely want to utilize the features. However, a 2.5mm (or 3/32) hex key (Allen wrench) is required to expose the jumpers. I would like to see AI provide this tool, or use conventional screws. Conventional screws are used to secure the XLR connectors, after all. Also, most gigging musicians do not carry a set of Allen wrenches in their toolbox. A screwdriver can be found at most any gig. Also, the manual provides no instructions on how to access the circuit board. Clearly, it’s not rocket science. However, the ribbon cable that connects the main circuit board to the patch bay is short. I would like to see the manual warn about this.

One last thing. It bugs me when manufactures omit a power switch. It is really inconvenient to plug and unplug the included wall wart to power up the unit when used separate from the Flex powered speaker cabinet. That said, the above-gripes are minor. If you need a two-channel preamp with XLR capability and flexible EQ options, the Flex Pre may just be your ticket.

1-5 (unacceptable to impeccable)

On-bench

SONIC PROFILE:

Low: Awesome control; can get big, but does the tight/full thing very well
Mids: Inherently neutral, natural-sounding mids
Highs: Stays rich up high; not in-your-face, but plenty of high-end content/detail