free web hosting | free hosting | Business WebSite Hosting | Free Website Submission | shopping cart | php hosting

Bender-120.ULT Amplifier Rebuild Kits Home Page

Info on the BENDER Rebuild Amplifiers.



The before unit, a stock Dynaco Stereo-120:

Photo by: Steven L. Bender
The BENDER-120.ULT - While there are no current plans for an article based on this design, or rebuild as I prefer to call it, but this is one great sounding transistor amplifer. While it looks pretty much like a stock Dynaco 120, it sounds quite different.

The ULT based Rebuild Kits for the Bender-120.ULT Amplifiers have been shipping since the 4th Quarter of 1999. Just to clarify one point, the ULT Rebuild Kits consist of only small parts, such as: new transistors, and proprietary semiconductors, new capacitors, new resistors, diodes, new RCA jacks, new lengths of wire and insulation, and so the ULT Rebuild Kit does not include a vintage Dynaco Amplifier. In most cases, the person interested in performing such a rebuild, already has a working, or possibly a non-working Dynaco Stereo 120 sitting on a shelf.

Since I love the sound of tubes, it may seem peculiar that I would concentrate much time and effort to a rather outdated transistor amplifier in this way, and still retain it as a transistor amplifier !!!

Well, since the introduction of the first transistor audio power amplifiers for home reproduction in the 1960-1965 era, people have spent thousands if not millions of hours listening to and debating whether tubes or transistors sound better... Audiophiles have also spent millions of dollars upgrading to get better sound during the next three decades.

Some Transistor History:

Transistors have certainly matured during the past forty years since the first devices. The early transistors were very crude, this so called Point Contact type fabricated using Germanium [ Ge ]. These were quite pathetic devices, soon replaced by Alloy-Junction devices which could be made of either Germanium or Silicon, this I would call these the First Generation of widely usable power devices, which included the first TO-3 cased power transistors. Closely following Alloy Junction was the Germanium Drift-Field Transistor. The Drift-Field Transistor power devices actually had some advantages, including: high beta, very good beta linearity across wide current ranges, and temperature stability up to 80 degrees Centigrade, along with extended frequency response capability. Despite the extended frequency response capability, these devices were still fragile and had limited power handling capabilities. Yes, some fairly good low power amplifiers were made in the late 1950s / early 1960s, using the 2N2147 Germanium Drift-Field power devices (some early transistorized Heathkits come to mind, like the AD-21 integrated amplifier). But, on the other hand, Silicon [ Si ] devices had some design advantages which outweighed the disadvantage of needing to derate the devices when operated above room temperature. As a result, the Silicon devices flourished while Germanium almost disappeared from the scene by the mid-1960s.

The early Silicon Homotaxial 2N3055 an NPN device, is actually a Silicon Diffused Junction transistor, which became quite popular around 1962, and was used as the silicon output device of choice, in many early transistor power amplifiers, including the Dynaco Stereo 120, and the AR Integrated Amplifier. Needless to say, Silicon manufacturing improved from 1961 to 1966, which gave us the: Double Diffused Epitaxial Mesa, Overlay Epitaxial Planar, Double Diffused Epitaxial Planar, Triple Diffused Junction, and Triple Diffused Epitaxial Base Planar Transistors. So by the mid 1960s we had this whole universe of silicon transistors - high gain, low power devices encapsulated in plastic cases, or small TO-18 metal cans, and medium power types housed in TO-5, TO-39, and TO-66 metal cases, and high power types housed in the larger TO-3 cases. As these Silicon devices were available in both NPN and PNP polarities designers had many more choices to choose from.

Despite all those choices, since the mid-1960's not much has actually been improved in terms of better sounding transistors. Confusion ensued, and The Tubes Versus Transistors Debate has had advocates on both sides of the fence going into its forth decade now... Reviewing Silicon Transistor's maximum Ampere Ratings and Power Dissipation capabilities, I've noticed that while both have increased - but really, only 200 to 250 percent for TO-3 style transistors during the last 39 years. If we consider the progression of Double Diffused Mesa and Planar Transistors as the 2nd Generation; then Triple Diffused Junction transistors becomes the 3rd Generation; then the Triple Diffused Epitaxial Base Planars, along with the not too successful - Ring Emitter Transistors comprise the 4th Generation. Now, as we finally approach the Millennium - Silicon devices have now made it to: The 5th Generation, and the net result is: finally - Better Sound from - the UltraLinear Transistors as used in The Bender-120.ULT amplifier rebuild.

Progress ? Yes ! But no matter how carefully a designer attempts to control the topology, it is a physical fact that across all the types of NPN-PNP Topologies as cited above, the complementary designs using both NPN and PNP device types - simply treat the NPN and PNP sides as equal and exactly opposite... Which they are not. Not even close. The resulting designs, starting in the middle to late 1970's abound. This is a case of multiple stages having a particular type of non-linear functionality, feeding several additional similarly non-linear circuits within the overall topology... That is why, most power amplifiers, even MOSFET designs, made from around 1975 through the later 1990's, tend to all sound quite similar. No longer do we have the early "Transistors Sound" that people complained about, but a kind of flat, lifeless, sterile, and smeared sound. And that is why I, as a designer, make the design choice to avoid the NPN / PNP pairs in power output stages, as I believe they tend to smear the sound. Why ? Well, the feedback loop and feedback signal, cannot differentiate between the devices in the node to node loop, and the complementary loop devices will have different characteristics, and thus the effect will be unequal for different paths on the + and - traveling cycles of the signal. This, especially where large amounts of overall loop feedback < Negative Feedback > is used, or where significant power is dissipated or delivered in that part of the circuit will cause smearing problems.

Lastly, I must add, that the real-world loads we use < Speakers > are very complex devices from an R / L / C aspect. The crossover components, and speaker voice coils Inductance L and Capacitance C aspects - of real world loads on a complex plane, are probably responsible for many of the sonic differences heard for a given amplifier, when played in different audiophile systems. Thus the feedback loop, also comes into play, as the fed back signal is affected by the interaction with the crossover/drivers which are far from a simple resistance at all audio frequencies. So again, to belabor the point, an output stage using the same polarity transistors is likely to fare better under these conditions, than the usual, complementary-symmetry circuits found in various Hafler's, Rowland's, Krell's, and Mark Levinson's, and most other designs seen sing the early 1980's.

Now More on My Rebuild:

The Dynaco Stereo-120 circuit is a true Minimalist Design, in essence it looks like a big Voltage Gain transistor followed by a big Current Gain transistor. It could not be much simpler - and still work. To a point, simpler is often better, and as it turns out, only the rather primitive semiconductors used back in the early 1960's held back the true sonic purity of this design.

The original Dynaco Stereo 120 incorporated three unique design features which are improved upon in the Bender-120.ULT rebuild. The First Feature is the highly regulated power supply, which is turned from a low impedance supply, at normal current levels to a high impedance, very current limited deregulated supply in the near short-circuit condition. While this limits the overall power capability of the amplifier into very low impedance loads, or an accidental short circuit, this feature protects the amplifier channels from that extreme misuse condition. This voltage regulated, and current limited power supply feature is retained in the rebuild, while the maximum power output of the original design is increased from 60 Watts rms per channel into 8 ohm or 4 ohm loads, to 70 Watts rms per channel, and the rebuild increases the Regulator's available current capability by 40%, without compromising system reliability, due to the rebuild's increased Margin in the output stage at all audio frequencies. Still, when called upon to deliver excessive current, which would compromise the amplifier circuits {as in the case of a short circuit} the Regulator will immediately cutback, delivering less than 1/4 Ampere into an amplifier threatening load, or short circuit.

Let me define this term Margin. In amplifier power output testing, the standard is defined by FCC Precondition Rule a two Stage Test, which first powers the amplifier using a 1 Khz. test signal level into large test resistors of a specific resistance, such that the power applied to the resistors is 1/3rd rated power for a period of 30 minutes, followed by a Full Rated Power test for which lasts for 5 minutes. After FCC Pre-Condition conditions, a properly working stock Dynaco Stereo 120, is actually able to sustain 53 Watts per channel into an 8 ohm load resistor at the clipping point. In addition the whole amp has gone from room temperature of 23 degrees Celsius to a quite hot 88 degrees Celsius after the FCC Pre-Condition Test. That this design could not quite meet its power spec is due to the more stringent Federally mandated FCC power output testing procedure, which came about some twelve years after the Stereo 120 went into production. It is also due to the limitations of the early Homotaxial Silicon output transistors, with their large value of Collector-Emitter Saturation Voltage and relatively limited power dissipation [compared to the best of modern day power transistors].

Here are some paragraphs provided by several transistor manufacturers quoted from their Design Manuals - dealing with the specification of: Transistors Active Region Safe Operating Area -

" There are two limitations on the power handling ability of a transistor: average junction temperature and second breakdown. Safe operating area curves indicate Ic - Vce limits of the transistor that must be observed for reliable operation; i.e., the transistor must not be subjected to greater dissipation than the curves indicate." " The data of figure 13 < a thermal de-rating curve of Power Dissipation [ Pd in Watts ] versus Case Temperature from 0 to 200 C > is based on Tc = 25 C < room temperature 78 degrees Fahrenheit >. Tj(pk) is variable depending on conditions < applied voltage, current, and power level >." " Second breakdown pulse limits are valid for duty cycles to 10% but must be de-rated for temperature according to Figure 1. At high case temperatures thermal limitations will reduce the power that can be handled to values less than the limitations imposed by second breakdown. "

Translated to more meaningful terms, the above techno-speak means that the silicon chip inside that transistor must be de-rated from its nominal power ratings when it is operated above room temperature [ above 25 C ]. This type of power versus temperature problem is quite typical of all the Silicon based transistors made during the past forty years.

While the design of the original Silicon output devices in the Dynaco Stereo 120 was adequate for operation at room temperature, when the amp is operating, due to very limited heat sinking, the temperature of the output transistors rises. At the very elevated operating temperatures the nominal power dissipation under loud music peaks, is reduced to about 70 Watts at an internal transistor temperature of 88 to 90 degrees C. And as such, its Operating Point is causing too much power to be dissipated inside the transistor chip as heat - with the result that the device margin is now near Zero, and the probability of transistor failure skyrockets. While normal operation isn't likely to heat the entire chassis of the amplifier up to that 88 degrees C. temperature, the internal temperature of the output transistors do repeatedly thermally cycle to, and beyond that temperature with the music [tens of thousands of thermal cycles over time]. So the older style transistors will suffer thermal stress diffusion cracks in the dice, and over periods of time will degrade.

The rebuilt Bender-120.ULT under the same conditions and tests will reach almost the same operating temperature, but, at that same thermal point, it will be able to provide 70 Watts per channel of output power, at the clipping point, under those same test conditions into the same 8 ohm load. Under that condition, the new output transistors still have a Device Margin over 100% at the 88 to 90 degree C. thermal point. At this elevated thermal point, each power transistor can still dissipate over 150 Watts, which is more than twice the power inside the output circuit, on loud music peaks. So, under the worst test conditions the Bender-120.ULT still retains over 100% margin - worst case, long term, under thermally cycled temperatures, it will not even reach 50% of rated power dissipation [PD.] and, as a benefit, the UltraLinear Transistors will not degrade as the original devices did.

The Second Feature of the original Dynaco Design - is the Class B current gain output circuit with bleeder currents. The output transistors are not operated in the usual class AB1 or AB2, but essentially operate without quiescent current, but instead, the current is forced through large bleeder resistors. The Rebuild replaces the original 7 Watt resistors with new 10 Watt units, often the 35 year old, often overheated 7 Watt cement resistors develop cracks, burn marks, and as such, they are compromised. As a result, the old resistors value can wander when the resistor is hot, another potential failure mode affecting the original overtaxed, and under-rated output transistors. Use of new 10 Watt resistors, rated 30 per cent higher, increases the reliability factor. As for the class B operation, strange as it may seem, and contrary to what one might think... in this simple circuit, there is absolutely no sonic penalty. Due to this simplicity of design, the use of a zero quiescent current design gaves much less of a performance hit than would be found in an otherwise more complicated modern multi-stage balanced complementary-symmetry output stage.

The Third Feature is the NPN-NPN quasi-complementary output stage with capacitor coupling. The simpler, and cheaper, alternative, often found, is unregulated plus and minus power supplies, and no output capacitor ( a configuration that was found in the old AR Integrated Amplifier, and 98% of amplifiers designed after 1980 ). The latter design was probably a cost-cutting decision, that left a serious long term design flaw in the unit. The AR Amplifier was also a 1960s design, which overstressed its early silicon devices, despite the presence of a speaker fuse, the lack of Margin in the output stage design, and the lack of an output protection DC sensing relay circuit, or an output capacitor is a flaw that would often lead to amplifier catastrophic failures. It has been documented, that some of these failures did result in fires inside of the speaker systems attached to these early AR Integrated Amplifiers, and some of The AR Receivers... Other similarly designed units { too numerous to list } have been known as woofer blowing, self-destructing... fireboxes... Errr. Amplifiers.

One could cite dozens of pages - of transistor data from the manufacturers data books, showing extreme asymmetries between the opposing NPN / PNP matching complementary devices. One can only imagine how the music suffers in the hundreds of more complicated designs which simply ignore these very real device differences ???

Now, getting back to the basics of my Rebuild Process, the primary improvement in The Bender-120.ULT { Level 1 / Level 2 / Level 3 } rebuilds, is the use of vastly improved semiconductors in the Current Gain Output Stage, with some improved capacitors in the signal path. Level 1 is no longer offered, as Level 2 isn't much more costly to implement and does offer some sonic improvement. The Level 2 and Level 3 rebuilds retain the original three large standup caps - the 3,300 uF / 100 Volt regulated power supply reservoir, and the two 3,300 uF / 50 Volt output caps. Often these caps are still within their original specifications, thirty-five years later. However, some are not. I've seen two such original caps fail completely in otherwise working amps during the past two years. Unfortunately, it requires an ESR < Equivalent Series Resistance > meter and also a Capacitance or LCR meter to properly test these caps, there is no way to do this without both of these pieces of test equipment.

Since the audio signal passes through the two output caps < which are in the AC feedback loop, and are present to protect the speaker from catastrophic failures > it is certainly necessary that these caps are both of the proper value, and it is essential that they not fail. Failing shorted would take the speaker out, failing open - the sound stops. That they have their nominal value is essential for proper bass response. And the ESR value is more likely to affect the high end of the audio spectrum sound and distortion levels.

The Bender-120.ULT { Level 3a } rebuild changes out the two standup power supply caps, and the amplifier output capacitors. Depending on the exact variation, different sets of caps may be used, in any case, they will be a set of modern production units with an extremely low ESR. The output caps chosen, at 100,000 Hz. these caps giving an ESR meter reading of 0.01 ohm, barely above the residual reading of the most capable laboratory reference meters. Higher values of capacitance insure a lower end primary rolloff pole from a nominal 8 Hz. to under 2 Hz. while at the same time increasing system reliability by having recently manufactured devices in the signal path, instead of devices which were stressed during twenty-five to thirty-five years of use. At Level 2 / Level 3 / Level 3a, the Servo Regulator Transistor is changed, as well as some additional resistors in both the Regulator Circuit, and in the Audio Amplifier signal path and in the Output Stage. Several Audiophile tweaks are also applied to the original design to fix the possible known problems occurring in older equipment designs. Gold plated RCA jacks, and heavy gauge grounding wires are installed in critical ground path areas, are retrofitted in both the Servo Regulator high current paths, as well as in the Audio Output circuitry. Problems in either of these areas could incur significant sonic penalties if high circuit resistance paths exist. Silver Solder is used to insure clean connections are made.

The Bender-120.ULT { Level 3 and Level 3a } rebuild now has the new OPTIONS 1/2/3, as of the 2nd Quarter of 2001. This is the first revision since the 1999 introduction of the Level 3 / 3a, when it was in Beta Test. Overall, this improves unit to unit consistency in several areas, and the overall operation of the input stage, lowers the global feedback, and provides improved performance of the Class A Driver Stage. Other changes decrease power supply residual ripple and harmonics; while improving channel to channel separation, while still retaining the Super Matched Pair of ULT Output Transistors. These { Level 3 / Level 3a } ULT Output Devices are specifically matched at the factory for higher current gain levels. Then, once I get them, I again match them, but at a very low current level, making them a two point match - Super Matched Pair. These SMP ULT Transistors are the result of much thought, and result in vastly improved sonic detail at low volume levels, while the Ultra-Linear Transistor characteristics at high current levels are responsible for the overall linear characteristic of the resulting sonics of the { Level 3a } rebuild amplifier's sound.

This combination of enhancements, provides truly audible improvements; while other Audiophile Tweaks which increase expense, but provide no sonic improvements, have not been included.

The resulting sound is unlike any other transistor amplifier that I know of. It almost has no sound of its own. Clean. Neutral. Sweet. It simply lets you hear more of the music - like tubes.

For those who know, many transistor amplifiers have great bass. However, this overly heavy "thundering thumpy" bass may be a form of transistor distortion. If you're ever heard the Stock Dynaco or Crown transistor amps from the 1960's and 70's; the bass seems to over-power the rest of the sound. My Bender-120.ULT rebuild leaves a more refined, and appropriate sounding bass end in sealed box, ported, and electrostatic speakers. When there is no deep bass, it might seem like the amp is bass deficient, but this is an illusion.

Likewise, The Bender-120.ULT rebuild doesn't lose low level details at very low volumes, a feature which is extraordinarily clear and detailed in the Level 3a design. Smear and loss of low level details is a common failing in almost all transistor based amps. The Bender-120.ULT also doesn't get shrill or harsh on the wide dynamic range of the best recordings. There are no clicks or chirps as the amp enters or recovers from clipping. It is the only really pleasant sounding all-transistor amplifier I know of. The { Level 3 } refinements gets it yet another few percent towards that level of clarity which makes it sound like a much more powerful version of my Single Ended Triode BENDER-3v. Almost. Enough Said.

Consideration is also being given to producing UltraLinear upgrades for other vintage Integrated and Power Amplifiers, and possibly one or two selected Receivers, specifically, the mid-1960's AR Amplifier, and the Heathkit AA-15 Integrated Amplifier. Later this year I'll be checking whether it is feasible to bring these vintage units into the next Millenium, sound-wise.

Click here to view photos of The BENDER-120.ULT Level 1 Amplifier

While the logistics of making up kits for tube rebuilds was way beyond my expectations, rebuild Kits of parts for the Bender-120.ULT will be made available for interested DIY builders.

This page is under construction ( Like software, the job is never done!....)
Prior Update - 05-02-2001 3:30 PM.
Current Revision - November 2nd, 2010 @ 9:50 pm.

Comments/Suggestions/Critiques/Questions to:
Email Alternate email: Email

This page, and all contents, are Copyright(C) 1997 ... 2010 by:
Steven L. Bender

Return to Steven L. Bender's HOME PAGE

Return to Top of Page