DaviSound TB-3 On-Line Manual
Printed text excerpts 1982 -1994 by MCP-DaviSound   All Text Copyright 1998-2007 by DaviSound
Original TB-3 "Inner Tube" Styling circa 1982

NEW TB-3 Styling 2006
Now Standard In 3RU Size and Popular "WG" Format

The DaviSound TB-3 "Inner Tube" Compressor
Operations Instructions and Applications Manual

"Rules of Thumb (and fingers) for operating the DaviSound TB-3"


The prime rule is to quickly acquire AWARENESS ... awareness that UNLIKE MANY OTHER pieces of equipment manufactured with limited scope, the TB-3 offers YOU, the professional operator, "LOOSE REIGNS" for control! This also means , that due to these "loose reigns" purposely designed into the TB-3 for broad applications at the discrimination of professionals, it may, therefore, take a little time to learn all of the "nuances" of the controls' relationships to each other.
But you will enjoy the experience!

The DRIVE Control adjusts the amount of incoming signal level directed to the tube stage. Any setting, other than full clockwise, is actually an ATTENUATION of the input signal with the difference being made up, and amplified further, by the tube stage and succeeding stages.

Settings for this control depend upon the exact tube utilized but "ballpark" settings are about 9 O'Clock to 10 O'Clock for unity gain of incoming signals levels. Of course, the whole purpose of this control is for operator discretion as to the resulting "sound character" it provides, especially when used in conjunction with the "DISTORTION CONTROL", and RECOVER CONTROL, so it's setting is a primarily a matter of desired effect. Increasing the drive results in increased THD (largely 2nd harmonic) of the tube (see more details in "Test Tube" proof of performance section below).

The DRIVE control has no effect, of course, when the tube stage is switched out as it is by-passed as well.

The "DISTORTION CONTROL" is a small knob located just below the tube port hole in the center of the control panel. This control allows you to lower the "soft clipping" threshold of the tube by adjusting the tube's gain bias voltage. Here again, it is provided for application at the discretion of the operator and, as such, there is no guideline other than the fact that the FURTHER CLOCKWISE it is turned, the lower the peak threshold and greater the distortion factor of the tube stage. At FULL COUNTERCLOCKWISE setting, the tube is operating with MAXIMUM HEADROOM and with very little audible THD.

At the 12 O'Clock pointer setting, the distortion control adjust the tube plate voltage to about 1/2 the availabe power supply voltage and this is, typically, the setting to begin easily, audibly detecting the added 2nd harmonic content to the incoming signal. More on this control in the "Test Tube" section below.

The "COMPRESS CONTROL" is the MASTER CONTROL of the "DaviSmart" Optical Compressor designs!
It is the "smart" control which adjusts all compression parameters simultaneously, all of which are interdependent upon the other, as gauged by the sampling of the incoming signal!

A full description is provided in the "Overview" section.

The "RECOVER CONTROL" adjusts the ACTIVE GAIN of the final amplifier to +20db. With certain line level input conditions, it is therefore possible to overload subsequent stages in a system if this control is misused! This is an abnormally high amount of user gain adjust for any line level amp, or processor, but it is provided for use in unusual input conditions.

The normal position for this control is about 9 O' Clock (barely "cracked") as this is the UNITY gain setting. Too high settings will degrade the signal to noise ratio of both the TB-3 tube stage and the signal chain ahead of the unit. Of course, this control is also useful, in conjunction with the other controls, for raising a -10dBu signal level to the traditional "pro level" of +4dBu.

The "MIC CONTROL" provides continuously variable active gain control, from 0 to about + 45db, to the built-in, differentially balanced, DaviSound "MIC-ALL" microphone preamps. A UNIQUE FEATURE OF THE TB-3 DESIGN IS THAT THE MICROPHONE STAGE IS ACTIVELY MIXED WITH THE LINE INPUT STAGE ALLOWING BOTH TO BE USED SIMULTANEOUSLY IF DESIRED! If this is not desired, simply turn down the mic gain control when not in use.
The line input may be left floating when not in use with no degradation to S/N ratio.

Suggested first time set-up procedure reference settings for the TB-3 ...

To begin getting to know the TB-3 it is suggested that you start with the most basic, neutral signal chain application possible. This will be with a line level, high quality music program input signal applied via the line input jack and controls set as follows-

All pots fully CCW (off) EXCEPT the "Recover" control, which is set to 9 O'Clock (about 1/4 open). The Tube switch is in the "out" position (Toggle handle LEFT). This places the amplifier in the most direct signal mode with just one stage operating at about unity gain. This will be your most "neutral" through-signal setting. At this setting the amplifier will be passing the input signal in the most "uncolored" manner possible, with maximum headroom and maximum signal to noise ...
i.e. with the best overall dynamic range.

Next, to get the idea of the tube stage effect, flip the tube switch on (Toggle handle RIGHT). Since the drive control is at first fully CCW you will hear nothing. Adjust this control to around 9 O'Clock, or just slightly beyond (depending the exact type and condition of the tube utilized) , and the tube stage will now be operating at unity gain also. You can fine tune this for the exact volume match by flipping the tube in/out switch back and forth and listening carefully until the levels are matched as closely as possible as you rotate the "drive" control.

If you can get the levels adjusted fairly close to being perfectly matched, you will notice something very surprising. You will notice that, at this setting, there is very little discernible difference at all in the sound with the tube switched in! This may teach you something about the so-called "sound of tubes" that you had never realized.. .that being that a single-ended, Class-A vacuum tube circuit, even one operated with relatively low plate voltage and NO negative feedback, can sound very "neutral", at unity gain, when A/B compared at the same exact signal level with a transparent state-of-the-art solid state stage! The THD is there anytime the tube is switched in, it is just not as obvious as you might think with most source material.

Now, to hear some increasing subtle, musically related (even order) harmonic content added by the tube, slowly raise the drive control to about 10 or 11 O'Clock. You may want to lower the "recover control", or not as the case may be, to compensate for the volume increase. Depending on program content, this will give varying degrees of subtle "tube warmth" or "fatness" to the sound. You can continue to experiment by raising the "drive control", and lowering the "recover control" until the distortion content becomes overly noticable or objectionable.

You should learn from these early experiments that even in the so-called "starved plate" condition, the tube offers very "clean" performance at unity setting with very "useably clean" subtle overtones added just slightly as the control is advanced beyond 9 O'Clock and kept below 12 O'Clock!

You can further experiment with the "starved plate" and "space charge" bias effect by advancing the "distortion control" further CW while listening carefully (and optionally watching on a scope). The CW advancement of the distortion control varies the plate voltage with the moderate plate voltage supply reduced ever further as this control is increased further clockwise. While this, theoretically increases distortion by lowering headroom, it may, initially, LOWER distortion since it also, simultaneously lowers the tube amplfication factor. You will likely need to advance the recover control as the distortion control is increased further on because of this lowered gain effect.

This description, so far, should just begin to reveal what we meant early on about all the inherent "loose reigns" and the relative interactions of the various controls that you wil continually discover as you use and operate your TB-3 in the months ahead!

Microphone Input Testing ...
Next, you can experiment with the mixing effect of adding a microphone source. It is suggested that you begin with a high quality condenser microphone plugged up to the microphone input jack. With the mic gain trim control fully off (CCW), turn on the phantom power switch for the mic channel under test. This will be the switch just to the lower left (or right) of the tube port hole, center panel, as diagrammed above. When turned on, the respective yellow LED indicator will light just above the switch.

Now, slowly adjust the mic gain control to 9 O'Clock. Speak into the microphone. You should, then, hear your voice faithfully reproduced and mixed with the line input signal. You can blend the two by adjusting the mic gain slightly and adjusting the overall level with the "recover" gain control.

Next, you may want to begin experimenting with the compress control in relation to what you have just established. You will likely find that a setting of around 12 O'Clock is required for the Comp LED indicator to just begin flickering. This is a moderate compress/limit setting. Try advancing the control to 3 O'Clock, and finally fully CW, in order to experience the effects of the various settings.

At this point you will be just beginning to discover and learn the powerful processing possibilities, and capabilities, of the TB-3. If you are like most of us, you will be quite pleased, to say the least, and perhaps even somewhat amazed at what you will discover!

You will begin to thoroughly enjoy owning and using a processor of the highest quality, cleanest operation, and, likely, greatest versatility of any you have ever heard/experienced previously!

The TB-3 even amazes us every time we run a proof of performance on one! It is a MOST satisfying experience for us to realize what we have acheived with this design, just as it will be for you to experience the satisfaction of owning and operating this precision piece of equipment!

As explained often in the manual, the TB-3 offers a lot of EXTRA reserve gain and control capability, the extreme of which will likely only ever be used under extreme conditioning requirements. But, it is there for you when/if you need it. Most of the time, you will likely NEVER need to advance any of the pot controls, with the exception of the "Compress" control, beyond their 12 O'Clock settings! Sometimes the recover control may be useful at around 3 O'Clock and, ocasionally it may be needed to run "wide open". Most of the time, however, you will likely only need turn your comp control,and recover control, somewhere between 12 O'Clock and 3 O'Clock, depending of course on signal input content and level as well as the desired effect.

Hopefully, what we have documented here will get you well into the "ballpark". The rest is up to you!

The TB-3 "WG-3RU" with wine colored "PurpleHeart" top, enlarged cover port and Walnut face!

As of 2006, all "Inner Tube" Top Ports have now been standardized at an ideal diameter of 2 1/2 inches so as to better facilitate tube maintenance and removal when necessary.
This eliminates the necesssity for top cover removal.

The tube may be removed for replacement by carefully reaching inside the "firehole".

When removing a tube ...
First be certain that power is turned off, the power cord unplugged, and the unit is in cold state !

Then, grasp the top of the tube firmly, but gently, between the thumb and two fingers.
Next, with an easy, slight rocking motion ...
CAREFULLY work the tube back and forth while pulling just slightly until the pins come free.

And, all tube mounting boards are now "MPT style".

The new MPT block features a hardwood circuit board ...
with all circuit components mounted underneath ...
rigidly embedded in thermal epoxy.

The boards are mounted on one half inch rails for a slightly
elevated position above the cabinet floor.

The first ever MPT shown above ...
made from quilted African Mahogany for a touch of the "exotic" on the inside!

One of the features of the new TB-3 is a dedicated utility ground lug on the rear panel. This tie point, which is direct at the TB-3 power supply common, is NOT normally needed for interconnection since signal grounding is accomplished through pins 1 of the input/output jacks. But, it is provided for utility purposes for those who have their own unique ground system requirements or who may prefer star grounding with all shields lifted at one end.

One of the main side benefits of wooden rack cabinets is their inherent isolation from chassis ground loops and using this terminal to ground the unit to a rack cabinet would, of course, defeat this side benefit!


Since launching our first vacuum tube design in the 1980's, the TB-3 "INNER TUBE" has gone through several stages of evolution to it's current state of progression. A brief description of this evolution should be of interest and provide useful information to all owners/users.

Our first design was simply a traditional tube amplifier stage with the opto-isolator circuitry acting on stage gain for a true, tube replica of our exisiting TB-2, solid state, compressor/amp. The result was a processor that was considered "TOO CLEAN" by some of it's users! There was very little "tube sound" (tube generated even-order harmonics, a desired form of "musical" distortion) with that first unit. The tube stage operated at a rather low gain, with considerable headroom. Therefore, the tube stage was very clean. Also, the gain reduction stage was AHEAD of the tube stage and, thus, further "cleaned up" the "glass act" even more in terms of overdrive capabilitiy.

We've already mentioned something of the way tubes naturally overload and the resulting, inherent, compression which takes place when a tube is overdriven. WE CHANGED OUR DESIGN TO TAKE ADVANTAGE OF THESE PROPERTIES THEREBY GIVING "TUBE-LOVERS" JUST WHAT THEY WANT - THE "SOUND" OF TUBES!

Tubes, of course, WILL eventually clip and the signal peaks WILL eventually flatten when overdriven far enough. However, long before this happens, slight overdrive, to a simple Class-A single-ended stage will cause asymmetrical signal swing with a subtle "bend", or gentle, rounding, of the signal peaks as opposed to the immediate hard clipping tendencies inherent to overdriven transistors of any topology.

We have done several UNIQUE things, circuit-wise, to give tube-lovers their "warmth". Our unusual technique employs just a single triode stage, one half of a dual triode per channel. We originally used the ubiquitous 7025 (12AX7) . But, in recent years, the earlier generation of larger octal tubes, the high gain 6SL7, and lower gain/higher drive current 6SN7, have once again become readily available, brand new, from Russia and China. These tubes "seem" to offer more of the aesthetic charm of the forties and fifties with a "tad" more inherent, "tubish warmth" than their miniature descendants.

They are also more ruggedly built and typically provide lower noise than a run of the mill 12AX7, especially being more consistently immune to microphonics! So, early in 2006, largely because of the availability of these tubes, our new TB-3 "WG" - 3 RU was born in order to utilize these wonderful old bottles. Whether you prefer, "new old stock" or newly manuactured, tubes of this family now seem very plentiful at reasonable cost.

Our manipulative circuit techniques force the tube into rather early overload, with resulting asymetrical signal swing, thus generating more of the desired EVEN-ORDER harmonics early on. Furthermore, our manipulations actually SUPPRESS the "nasty" odd-order harmonics ( those that are most audible, non-musically related, and, thereby, distasteful to the human ear). The result is a "silky, sweetness" which enhances the music signal through harmonic colorations. The processor, as it is gradually overdriven, actually behaves like a musical instrument generating warm, musical overtones which are FREQUENCY INDEPENDENT (unlike digital synthesis) throughout the entire audio spectrum!

In other words, no matter what the digital synthesis mathematicians may like you to believe ... the ONLY way to get REAL tube compression is to use a REAL TUBE!

Stage by Stage ...

The front-end "DRIVE CONTROL" allows you to adjust the amount of tube overdrive, with the resulting harmonics and natural compression, by "pushing it" as much, or as little, as desired.

Next, the signal travels to the COMPRESS stage, which is adjusted by the COMPRESS control. This stage was placed AFTER the tube stage so that when the signal is "squashed" by normal compressor action, this action further emphasizes the harmonics (musical overtones) already generated by the tube prior to compression. All of this makes for a very versatile, fully "tube sounding", processor with a very large degree of user control.

The next stage is the final buffer amp which may be adjusted by the active, "RECOVER" control for either NEGATIVE gain or POSITIVE gain as required after compression and/or overdrive. While this control will adjust the amplifier to less than unity gain, it will NOT lower it below -3dB and therefore not function as a full off volume control. It is a TRUE amplfier stage active GAIN control set for partial gain adjustment to provide makeup gain from the "RECOVER" stage after compression.

What you see is what you get ...

The front panel "COMPRESS LED" shows the amount of compression being employed beginning with just a slight flicker and increasing to brighter, more sustained, display as the COMPRESS CONTROL is advanced.

Unlike comparator driven displays which are, in effect, simple A/D converters, this display is a TRUE ANALOG of the compress drive signal thus giving a better "feel" for the intensity of the amount of compression being employed. However, that being the case, it will not be highly visible from a far vantage point. It is advisable to adjust the control with an eye directly on the LED during preliminary adjustment. You will then be able to see peak gain reduction early on (a slight flicker on signal peaks), long before any compressor action is audible, while the COMPRESS CONTROL is being advanced.

The tube stage operates with about 15 dB of fixed gain and the recover stage operates with 20 dB of active, continuously variable gain as deecribed above. This allows a great leeway in user-defined gain structures providing unity signal levels at various tube drive settings while, also, providing make-up gain following compression.

"Test Tubes" ... ...

We employ a hand-picked, burned-in, SOVTEK (made in Russia) 6SN7 that tube enthusiasts have now heard so much about, and/or heard from, as of late.

Of course, any 6SN7 replacement may be used, though we think the Sovtek military grade is possibly of higher quality than the new tubes from other sources that we have tested. They are also, often, the lower priced!

A 6SL7 may also be used as a direct replacement for different performance "textures" within the same support circuitry. But, we feel the 6SN7 offers marginally better characteristics with our circuit structure and the widest variation of tone and THD as affected by our related controls.

Although the 6SL7 has a much higher "mu" (gain) rating, it may actually exhibit LOWER gain in our application in the TB-3 because of the lowered plate supply voltage and the 6SN7's ability to draw higher currents at lower voltages!

Proof Of Performance Notes ...
Here are a few notes excerpted from the documented audio proof of performance tests which were run on the tube circuit stage of our first edition, new TB-3 "WG" utilizing a new, Sovtek 6SN7 tube.

Test Set-Up -
DaviSound low distortion sinewave sweep generator set to 1 volt at 1KHz feeding Sperry AC RMS voltmeter to Velleman Oscilloscope via TB-3 line in/out. The compressor stage was switched off and the final stage set for unity gain with the recover control at 9 O'Clock.

Basic tube effects ...
Once tube stage gain was carefuly matched using fine tuning of the drive control and the tube in/out switch for comparison, discernible waveshape disortion was just barely detectable at unity drive setting with distortion control full CCW.

"MAX HEADROOM" test ...
Unity setting of the DRIVE control for this unit was right about 10 O'Clock. The maximum headroom test for this unit's tube stage only, was conducted with the DRIVE control at this unity setting. The "DISTORTION" control was, of course, set to maximum supply voltage at the full CCW setting.

Surprisingly, even in the so-called "starved plate", much lower than typical operating plate voltages for tubes, the stage did not begin to start full clipping (flat - topping) until +20dB output through the tube stage with the signal increase taking place at the signal generator (drive control remaining at "unity" position)! This is headroom performance almost as good (within 2 to 4 dB) as some professional, solid state devices being sold on the market today!

Carefully examining the scope, and returning the input signal back to 1 volt, further testing was documented. Increasing the drive control to 12 O'Clock and beyond showed only very slight, but some barely noticable change in waveshape. This was a very slight widening of the upper excursion of the wave which became fairly noticable at about the 2 O'Clock setting.

Advancing the drive control to just beyond 3 O'Clock showed a marked increase in the widening of the top wave excursion with a full rounding, almost flat-topping (soft clipping) by the time the full CW setting of the control was reached. The lower wave excursion was not noticably affected to any similar degree.

Next, the effect of the DISTORTION control was brought into the test. Advancing this control to the 12 O'Clock setting lowered the threshold of the soft clipping point (full rounding of the top slope) to the 3 O'Clock rotation setting of the drive control. Then, at this same 12 O'Clock DISTORTION CONTROL setting, advancing the DRIVE control to full CW position showed full clipping (flattening) of about one third the slope of the upper excursion ( the measured output, with recover set in it's uniity 9 O'Clock position as was done for all the tube tests, was +12 dB, just over 3 volts RMS, in this condition).

Advancing the THD control even further from this point to full CW setting showed only slightly earlier clipping point but mostly just reduced overall tube gain by almost exactly 6 dB.

Test summary ...
A/B comparisons with the tube in/out switch at "normal" settings may tell you something about your audio perception. There is ALWAYS siginificant, measurable (analyzer) THD generated by the tube circuit at ANY setting as long as it is in the path because it is running at its full gain ratio and current capacity (for the applied voltage) with no corrective feedback and working into an unusually heavy, forced load. However, it is unlikely that your ears will reveal much when in this condition during A/B comparisons! In fact, only very careful examination, by trained eyes, can detect much change to the sine wave shape on an oscilloscope at these low drive/low THD settings. This is why, for intentional higher THD generation by the tube stage, the distortion control will likely always be set at around 12 O'Clock to 3 O'Clock for more audible detection.

This brief test should also go a long way toward showing why tubes, when treated much kinder in terms of circuit support and supply voltage, are so desirable for high fidelity audio applications even in their simplest configurations! You can see that we have to go to great extremes to make them audibly "misbehave" at the same signal levels associated with interfacing solid state circuitry!

Tube maintenance ...
Tubes often last for years though some users advocate replacing them regularly, annually for example. When the tube becomes noisy (hissy, crackly etc.), or otherwise erratic in behavior, then it is certainly well past replacement time. Otherwise use it to your heart's (and wallet's) content!

While part of our old technique was to drive the filaments of the 7025 a bit harder than recommended, our new application of the 6SN7 utililizes a filament supply voltage that is actually slightly LESS, and thereby somewhat gentler, than suggested by factory specs! This fact, combined with the low plate supply voltages, will provide maximum tube life! In fact, you may NEVER have to buy another tube because of failure as it might just last indefinitely!

One "old-timer's trick" relating to vacuum tubes is to leave the gear on ALL the time or for as long a period as possible. This is because tubes are most stressed at power-up and their filaments most sensitive to damage at power-down. Many large studios, in the old days of tube dominance, NEVER turned their equipment off for this very reason. The longer they "cook" (heat), the longer they last and the more consistent they sound over time.

That "touch of glass" ...
Actually, utilizing a vacuum tube stage really DOES bring a unique form of processing to signals that, unquestionably, can NOT be truly simulated by either digital processors or some other analog means. That's why tubes have returned BIGTIME these days. Of course, they have never really left the pro audio scene when you consider tube condenser mics, direct boxes, and instrument amps, all of which have continued to play an important role in every studio's "sound character" portrayal all through the years.

There is only ONE, true way to get a real "tube sound" and that is by letting a real "live" piece of glass do the talking ("singing")! That is certainly what happens with your new DaviSound, TB-3, "InnerTube" compressor!

(All Text, Photographs, Descriptions and Terminology Copyright © 1998 - 2006 by DaviSound - All Rights Reserved.)

TECHINCAL SPECIFICS of the MP-3 "MasterPiece" Compressor Module

  • Overall amplifier gain (MP-3 Compress Module) +20 dB
  • Available Gain Reduction - over 30 dB
  • Channel matching characteristics +/- 1 dB
  • THD at 1 KHz, recover control full CCW, 20dB of compression - under .08%
  • THD at 10 dB compression - under .01%

  • The compression control settings and resulting affects are input signal dependent so there are no hard, fixed position indicators.

    However, were the control turned fully clockwise, the chain would be operating with over 30 dB of gain reduction!

    Typical MAXIMUM settings of the "comp" control would be somewhere around 12 O'Clock (knob pointer straight up) which typically results in 10 to 20 dB of gain reduction depending upon the input signal peak/transient content. The extra reserve of the "comp" control is there if you need it as described further below.

    While the on-board LED "comp indicator" is quite sufficient to provide an exact visual analog of the compressor chain in action, detailed results can be measured on any down chain VU meter in your system (recorder input meters for example). With a 1 KHz sine wave input to the TB-2 and the "recover control" set for approximately 10 dB of gain at about 12 O'Clock, as the comp control is advanced, you can observe the logarithmic gain reduction "action" characteristics, for a steady sine wave, displayed on the VU meter.

    If using models with the specially, custom ordered "gain reduction" VU meter display, the inherent decay characteristics (release) of the particular setting respective to signal input may be observed when the signal is abrubtly removed.

    Since early compressor designs reflected the applications and required use of their day, they often included "stereo cross linking". This was considered especially useful for one of their main purposes, disc mastering, in preventing cutter groove "worble".

    The "DaviSmart" principle does not subscribe to "side chain" or "stereo linking" application! We maintain that both channels of our dual unit can be better, properly, manually adjusted for optimum stereo application, when required, accurately processing stereo program material far better and more faithfully than if cross-linked in older, traditional "stereo" compressor application.

    Nowadays, when most "bottom end" and heavier bass transients are "centered" and common to both channels, "stereo" linking, where one channel's transients and signal weighting affects the other channel, is considered far too un-natural and undesirable to be of any merit when accurate tracking can be better performed, and matched, invididually and manually with the "DaviSmart" comp control and recover settings.


    The NATURE of the animal ...

    About COMPRESSION in general ...

    The more one understands one's tools, the better to use them for the job that is required. It is toward that end, of better understanding, and, thus, better application, that we offer an extensive general discussion of compression, along with specific information regarding our methods for acheiving it.

    Compression is just what the word itself implies. The dictionary defines COMPRESS as: "condense, constrict, densify, squeeze" ... to decrease, shorten, thicken. Well, in relation to audio dynamic range, ALL of these adjectives apply when audio compression is added in varying degrees!

    Most all audio compressors do, essentially, the same thing. The thing that varies is the specific way in which the various designers go about acheiving similar results.

    Compressors work on a GAIN REDUCTION principle. They usually employ a rather low-gain amplifier front-end with some form of automatic, variable gain control. This control is VOLTAGE VARIABLE. The more voltage this stage receives at it's control point, the LESS GAIN the amplifier produces. This voltage controlled amplifier, or voltage controlled attenuator, as the case may be, enables a compressor to be designed by sampling the signal passing through this stage, amplifiying and processing the sample, and sending it back to the control element to reduce the amplifier's gain.

    Now, the more this process is applied, and the greater the amount of compression drive that is utilized, then, the higher the RATIO of input signal to output signal becomes. In terms of decibels, a 10:1 ratio would imply that an input change of up to 10 db would result in an output change of only 1db .

    If enough gain reduction drive is employed, the compression ratio approaches infinity. This means that we have, at that point, amplified our sampled signal as far as we can and driven it back into our gain-reduced amp as much as we can, to the point where there is very little dynamic range left in our "squashed" signal to make any difference!

    All audio signals have a noise floor and a maximum peak ceiling. If we operate too close to the noise floor, our signal to noise suffers. If we operate too close to the peak ceiling, we're in danger of "bumping the ceiling" and CLIPPING our signal. "Clipping" is the point where an amplifier, or system, simply "runs out" of gain and headroom and the signal peaks begin to flatten causing very harsh distortion. The whole idea behind compressors, and peak limiters, is to reduce the dynamic range. This may be done either very slightly for practical reasons (as in peak limiting) or in very large degrees for a desired special effect.


    It's all relative ...

    No matter how one might TRY and isolate them, ALL of the parameters of audio compression are, actually, inherently very inter-related and inter-dependent. These parameters, such as RATIO, THRESHOLD, ATTACK AND RELEASE, are not only inter-dependent within themselves but, also, with the many,varied characteristics of the input signals as well!

    Our designs were among the first to take advantage of these inherent properties of signal dynamics. Our tests and developments in these areas allowed us to ultimately create a circuit design that would AUTOMATICALLY self-scale ALL of these parameters in the best relation to each other, based on the constantly changing qualities of the input signal itself, in a musical, sonically pleasing manner, as compression is applied with one, single control! Thus, the "smart" compressor was born!

    Our COMPRESS control, used in conjunction with the RECOVER control, essentially handles ALL of the parameters required based on the input signal sample. As the COMPRESS control is advanced from "full-off" to further on, in clockwise fashion, it causes both a lower THRESHOLD of signal action and a higher RATIO of dynamic range compression, simultaneously. As the OPTICAL ATTENUATOR reduces the gain of our signal as a result of this action, it becomes necessary to raise the overall "squashed" signal elsewhere in the chain to a higher AVERAGE level. This is accomplished with the RECOVER control which is actually a variable gain control to our final buffer amplifier stage.

    It is important to note EXACTLY what we are doing when we advance both these controls into action. We are, actually, reducing our PEAK level well below our CEILING level and, thereby at the same time, raising our system FLOOR where noise, distortions and signal "nasties" reside!

    To any audio purist, the question naturally arises, "Why would anyone want to deliberately LOWER dynamic range, the very thing that gives amplified audio it's approach to natural sound?"

    That's a VERY good question!

    In the early days of recording, it was desirable to employ limiters to avoid saturating tape, and amplifiers elsewhere in the chain, with limited headroom. Some people actually started to like the "sound" of this process even to the point of over-using it when it was not required for practical purposes.

    It would give an "up-front" presence to vocals and certain instruments in a busy mix causing them to sound "louder". Vocals and electric bass were, particularly, much easier to capture and record with fairly high compression ratios. Many artists began depending on the action of a certain type compressor/limiter for their mic technique and resulting "sound signature"! This became so "faddish" that it got to the point where a studio had to have a certain type compressor in the chain in order to record certain artists or the artists would go elsewhere!

    Processing of this type was carried to even greater extremes in broadcasting (and still is today!). AM broadcasters, in particular, almost always over compressed their signals for a variety of reasons. First, it gave them the maximum peak output power allowed by law since they could modulate their carrier signals right up to the very limit without overmodulating or "splattering".

    Next, since over-compressed signals sounded louder (actually harsh and fatiguing to many listeners!), they seemed to "jump out at you" in car radios which many program directors thought/think was/is desirable. In early Top-40 radio, it became "in-vogue" to see whose station could be the loudest on the dial! (and, sometimes, the most distorted along with it!)

    In the serious recording studio, compressors can be most useful tools since they allow, to some extent, a "set and forget" approach to recording certain tracks. Thus, they can become a second pair of hands with a lightning speed brain/reflex control system! On some instruments, such as electric lead guitar, they can be "overused" at times for good purpose by adding sustain and "bite".

    As we have mentioned already in some detail in other sections of this manual, scoped tests will show that tubes, when overdriven, have a natural compressor action inherent. They actually go into about a 10:2 compression ratio just below their clipping threshold. Guitarists have known this, instinctively, for years and have depended on this characteristic, along with the overloaded tubes' fat musical overtones, for their sustain and "sound".

    Compressors can be used in much the same way to help simulate or enhance this effect in the studio. The TB-3 allows for both! You can even get some amazing sustain on an acoustical guitar track in this manner and, if taken to the extreme, cause it to sound almost identical to an electric guitar played through a tube amp!

    Rules of thumb (and fingers) ...

    In general, when adding compression to an OVERALL MIX, or to an AUDIO PROGRAM, considering what we've covered so far, the LESS compression added, the MORE NATURAL things sound!

    Our compressors start out so subtle in their initial action, that you need a peak indicator, or subsequent chain stage VU meter, in order to SEE what they are doing before you can actually hear any effect at all.

    If OVERALL compression is desired for a general program mix, it is usually best to use a small amount so that signal peaks are limited but the more subtle sounds are unaffected. This can easily be accomplished with our "DaviSmart" designs by advancing the COMPRESS control to the point where the compress indicator LED just begins to flicker on signal peaks. If more compression is desired, the control might be advanced to the point where compression just begins to become audible and, then, backed off slightly.

    On the other hand, for isolated instruments and voices when laying tracks in the studio, higher compression settings may be desired. Electric bass is almost always "tightened up" with a fairly high compression setting. Vocals usually require moderate compression settings varying, of course, from one performer and performance to the next.

    Action ...

    Remember the little old lady who always drove thirty miles an hour no matter where she went but insisted that she have the horsepower under the hood just in case she needed to get from thirty to sixty in just a couple of seconds?

    Our COMPRESS control is designed much the same way.

    It begins at a very high signal "threshold" with the incoming signal and initiates about a 2:1 compress action. ATTACK and RELEASE are always automatically self-scaled according to the previous parameter settings AND the required handling of the musical dynamics present in the signal itself.

    The COMPRESS control reacts VERY QUICKLY as it is slightly advanced. At about 8 or 9 O'Clock, things begin changing rapidly. The ratio quickly increases to about 5:1 and this, in turn, lowers the threshold of operation. At around 10 to 11 O'Clock the action of the control begins to become more "hair trigger" (to provide all the built-in user leeway described earlier). By the time we get around to 3 O'Clock with a typical incoming signal, we are approaching infinity as our signal has become so "squeezed" that there is very little difference in the sound as the control is advanced to the extreme setting.
    (Actual control positions will vary from above depending upon the level and dynamics of the source material.)

    So, like the car that goes from zero to ninety in one swift press of the accelerator, our single control takes you from zero to infinity in compression in less than one full rotation. This can sometimes make fine adjustment a bit tricky but, with a little practice, and with the help of the LED compress indicator, operation of a "DaviSmart" Optical Compressor quickly becomes second nature!

    You'll find most of the applications require compress settings well below 12 O'Clock. Actually most of the required settings will range between 8 O'Clock and 11 O'Clock; but, like the extra, reserve automobile horsepower, the extra compression action is available from this one control for those occasions when you might want to apply it.

    So, you can see, there is enough leeway designed in to allow you to OVER -use the compressor easily. The design is, intentionally, NOT FOOLproof! It has to be applied discriminately, and wisely, but such should be the case with any professional tool. The point of all this precaution is simply this: NO other product that you may be used to offers you as much "action" from it's controls as the TB-3 COMPRESS control places at your disposal. So, get used to a lot more "action" from within one rotation and learn to use it judiciously!

    "Garbage IN... Garbage OUT" ...

    That old saying takes on an emphatic meaning in regard to audio compression. Actually, it becomes, "Garbage in, ever INCREASING garbage out!". This is because, if there are is "dirt" lurking anywhere within the signal, compression will certainly bring it up to where it can be heard or "felt".

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

    Live Sound / Broadcast Processing Application

    The TB-3 design can also be used to advantage in program processing application for live sound and for broadcasting. However, it is in these applications where the aforementioned extra compression action provided must be most carefully taken into consideration. For example, it would be disasterous for a program director, or station engineer, to put one of our compressors on-line and simply turn the controls "wide-open" as they sometimes do with other products! But, operating within the 8 O'Clock to 10 O'Clock setting range of the COMPRESS control will yield an excellent, high quality broadcast program "processor".

    Likewise, using a TB-3 for live sound, ahead of the final amplifier stack, will allow driving the system at a much higher AVERAGE level without fear of peak clipping. This is easily set up by turning the compress control up to the point where the LED indicator is just flashing on peaks. This will normally be with the control just "cracked", somewhere between 8 O'Clock and 9 O'Clock.

    Once again, going too far beyond this setting on full program material will result in increasingly un-natural sounds as the distortion floor of the overall program is raised to near peak level by over-compression. Of course, all competent professional sound engineers will be aware of this and would detect any audible artifacts of such a setting long before their audience.

    Integrity and Security ...

    We've gone to great lengths in this manual to emphasize the potential for MIS-use since no other compressor design, that we know of, has ever offered the user anywhere near as much processing from one single control as does the "DaviSmart" Perhaps, since we've considered so much "negative" with all the precautions, we should dwell, in closing, on some of the many, numerous POSITIVES you can expect from your unit!

    A lot of this "positive" will be apparent as soon as your ears respond to the signals passing through your unit! The silky, transparent compression action, when properly adjusted, will naturally speak for itself. You will hear no undesirable compression artifacts. One reason we spend so much time going over the powerful COMPRESS control action, is because the novice user will often "crank" the control until he/she hears some "hard compression". At this point, the TB-3 is probably OVER compressing, especially on mixed program material! This is because it does it's thing largely UNNOTICED in a very transparent fashion when properly adjusted and applied!

    Part of the FUN of trying out your new unit is to watch the gain reduction and compress action taking place on a downchain VU meter long BEFORE you can HEAR anything like traditional "compressor action"!

    As with all DaviSound products, you can be assured of the most knowledgable and careful circuit designs, component selection, application and circuit construction techniques to guarantee and assure signal integrity. You would expect no less when entrustuing your archive quality material to any processor and neither would we.
    (please refer to our "Design Philosophy" page for more information in this regard).

    By nature, all compressors, or any processor that alters the incoming signal in some fashion, are inherent signal "colorers" of a sort. To fully assure yourself of the integrity of a "DaviSmart" unit, carefully examine a quality sine wave through-signal at various levels and settings on an oscilloscope. YOU WILL NOTE EXCELLENT SINE WAVE SHAPE AT ANY SIGNAL LEVEL, FREQUENCY SETTING, OR COMPRESSION LEVEL. THERE IS VERY LITTLE DISTORTION EVEN AT DELIBERATELY EXTREME LEVEL SETTINGS!

    Our current design was inspired, initially in the early days, by the infamous "LA-2A" of the "old days". But, we took the concept so much further and, then, married it to state-of-the-art technology opto-isolator componentry. The result is, we think, the best multi-purpose audio compressor ever designed with "speed of light" attack time and moderate, signal dependent, variable decay time.

    After you have lived with your compressor and truly come to know it's capabilities, you will never want to be without one, or several (some use them on every track of a mulit-track format!) during any recording session ever again!

    Additional written specifics about your unit are included inside the box at shipping.

    Thank you for owning a DaviSound TB-3!

    Please let us hear from you regarding your application of your unit!

    Back to ..

    DaviSound Tool Boxes
    DaviSound Main Menu