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

Steven L. Bender's Home Page, with Information on:

The BENDER line of Rebuild Amplifiers.



I actually have been interested and involved in Audio design, for more than 20 years. Back twenty years ago, in 1977, as a Graduate Student in Audiology, at Columbia University's Graduate School: Teachers College, where I studied, at the same time I was inventing The Hearing Audio System, ( HAS ). This HAS unit was the first of a series of pseudo-binaural and binaural listening devices. The various HAS were battery powered portable listening units, special in their ultra-wide bandwidth and low distortion characteristics. Other hearing aid like devices, now perhaps more commonly known by the A.L.S. ( Alternative Listening Systems ) acronym, many are simply bare-bones imitations of my sophisticated HAS designs, from twenty years ago. My other HAS devices included: an Audio Loop System, and a Tinnitus Masker.

I didn't start to use tubes until about twelve years back, when I bought several Tube Preamps and Power Amplifiers, and learned the joys of the Tube's Sound.

On the design side, it wasn't until early 1996 when I started hearing and reading a great deal about the resurgence of tubes, that I started applying the technical info to what I was hearing. When I went to the HiFi Show back at the end of May, 1996 at the Waldorf-Astoria, in New York City, I found most of the systems were using TUBE amplifiers. The number of Single Ended Tube Amplifiers available, took me by surprise.

Yes, I had known about a resurgence of Tubes ( and the retro desire for Triodes ) and Single Ended Class A Amplifiers in general, but I didn't know it was this wide spread. I had been using a tube preamp and push-pull power amplifier for about twelve years, now.

So I decided to see what it was all about, by initiating my own S.E. Triode tube projects, to turn a Transistor Stereo Power Amplifiers into a Single Ended Tube Power Amp.

This led to several months of thinking, investigating, and checking schematics. Finally, I came up with several designs, some of which, were not likely to fit in the limited physical space I had alotted.

This problem of limited space was the first problem to be tackled; - by choosing a rather compact transistor power amplifier to serve as the chassis, I had to make some tough decisions that later affected the final design. The Dynaco Stereo-120 chassis was chosen for several reasons.

The before unit, a stock Dynaco Stereo-120:

Photo by: Steven L. Bender

Three reasons that I chose this old amplifier:

#1 - A large number of units were produced and are out in the field ( > 100,000 ).

#2 - I wanted to use a Switching Power Supply ( SPS ) to power the Tube Filaments, and SPS power supplies give off RFI interference, requiring extra shielding. The plus here is that the Dynaco Stereo-120 chassis is a big chunk of steel, with a steel cage.

#3 - The well known fact is - that a rather large number of those aging Stereo-120 units ( now 20-35 years old ) have been failing, many, due to the Standup Capacitors failing.

There are five "big" stand-up computer grade caps in the original Dynaco design, and many thousands of these caps are failing. The most common stand-up computer grade caps that I saw - which had failed, were those Made in Scotland (ones with the Green outer insulation) these are usually found in the later production units, those with the "TIP" driver transistor upgrade, with most of these caps bearing date codes in 1973, 74, and 75. For those trivia experts or interested amateurs, often, the different sources were denoted by the color of the outer insulation present on these caps.

As can be seen in the above photo, these five stand-up computer grade caps were actually sourced from over a half a dozen different manufacturers, all over the globe.

In general, USA (Blue-Mallory or Clear-Sangamd), Japan (Clear), Great Britain (Gold or Clear-TCC), and Scotland (Green), there are probably other manufacturers, that I hadn't noted. It was rare to see two amps with the same complement of caps, or all of one color.

The unit photographed, an early unit, with TO-5 driver transistors in finned heat sinks, is sporting has caps from all over the globe, and it is also somewhat unusual in that the Regulator and Left Channel PC Boards are Phenolic (brown) while the Right Channel is Glass-Epoxy (green).

Here are two failed caps, one that Vented and one that is Sprouting Goo:

Photo by: Steven L. Bender

It also deserves some mention that a problem that plagued many of these Dynaco amplifiers - that they were kits and the soldering by non-professionals was often substandard. I shudder when I think of the bad solder joints I've encountered in the various dozens of Stereo 120's I've opened.

Bad soldering will cause any circuit to fail, no matter how high the quality of the circuit and the parts.

One notable example concerns a Stereo-120 that I bought within the last year. As it had, a hum problem, the previous owner had fashioned a steel shield, to cover one corner of the power transformer, and this shield was grounded with a wire to the chassis. The problem, I found was that the speaker ground lead was simply soldered to itself, and was hanging off the binding post, making contact with absolutely nothing. Poor soldering to the extreme.

As there are over one-hundred twenty thousand Dynaco Stereo-120's currently out in the field, and all are now from twenty to thirty-five years old, it is likely that a large percentage are starting to fail, or have already failed. That being the case, I felt a large number of possible upgrades was possible, even desirable. So, my first three Tube Designs, could more accurately, be described as rebuilds of those older transistor amps.

By, re-using the chassis and power transformer and a few other parts, one saves a significant amount of money, and labor. Starting with a blank chassis is, of course, possible, but increases the cost, and decreases the likelyhood of a professional look at completion.

It also makes the labor-intensive time to build it, about twice as long. I spent a considerable amount of time, just insuring that the rebuild amps could actually be implemented in the limited physical space in the Dyna Chassis. So at this time, the first three designs became:

  • The BENDER-3v. a 4 Watt per channel Single Ended mode Stereo Power Amplifier, using three Paralleled Beam Pentodes in Triode Mode in each channel.
  • The BENDER-2pp. a 12.5 Watt per channel Push-Pull mode Stereo Power Amplifier, using a pair of Beam Pentodes in Pentode Mode in each channel.
  • The BENDER-1v. an 8 Watt per channel Single Ended mode Stereo Power Amplifier, using a single Directly Heated SV-811-10 Power Triode in each channel.
  • Stay Tuned...

    Pictures of the first three Amplifiers as they existed in September, 1996:

    These pictures will be updated once the final versions are available. Most likely these three will become published, as I have discussed the details with the Editors of the following magazines:

    Click here to view photos of the early version of the BENDER-3v. Amplifier

    The BENDER-2pp. - has appeared in Nuts & Volts Magazine. The design of this Push-Pull Pentode Stereo amplifier is detailed in the February, March, and April, 1977 issues - as a three part D.I.Y. project.

    Click here to view photos of the BENDER-2pp. Amplifier

    The BENDER-1v. - this might appear in some audio magazine, such as: Glass Audio, Sound Practices, or maybe the Newsletter of the New York Audio Society [ S/N ] or somewhere else, :-). But given the unknown situation with Svetlana and as the only source of these SV-811 and SV-572 tubes, which are perhaps in short supply as i write this, I might at some point switch to a design based upon the newly introduced Electro-Harmonix 7591 tube, which has previously been used mostly, in various Fisher Receivers of the mid-1960's, such as the 500B, 500C, 800C and so on, and maybe one or two gutar amps.

    Click here to view early photos of the BENDER-1v. Amplifier

    Design Details of my Tube Amp Rebuilds

    Input/Driver Stage

    The input stages in my minimalist designs is simple. In the Single Ended Designs, one Triode or a Triode connected Pentode comprises the gain stage. Then a Cathode Follower powers the output stage.

    That Cathode Followers may have gained a sort of bad reputation is probably undeserved. The Miller Effect is a fact or life in Tube Circuits, and it is a tube capacitance that causes a roll off in the usable frequency response. This effect becomes especially bad in Open Loop amplifiers ( those that use no Negative Feedback ).

    Using just a gain stage would lead to the problem of small signal decrease in gain, as frequency increases, which only gets worse for large signals.

    My original design for The BENDER-3v. used two input triodes paralleled driving the output stage. It exhibited a -6dB roll-off at 20Khz. for small signals, and that got worse, as you approached full power. On some speakers it sounded fine, but others, it was noticeably rolled off. That is why the Cathode Follower is used, it can pump significant current, without loosing its cool, and without continuously altering the frequency response, as frequency and level changes occur.

    The BENDER-2pp. Push-Pull Amp uses a dual Pentode/Triode tube, wired as two triodes, as the voltage gain stage and cathode follower type phase splitter. The output stage operates at less than Class A, as each output tube does mostly half of the signal waveform.

    Some overlap is desirable, so about 5mA of current biases the driver stage, which feeds the output stage.

    Bias Circuit

    The BENDER-3v. and BENDER-2pp. amps, using small output tubes use a Bias Circuit which is non-standard, combining some negative DC bias and Cathode Bias level and using a Cathode Resistor. This insures that proper operation will be maintained, at the expense of a small nominal amount of power.

    The AC output is adjusted using a pot in the Push-Pull design. The Bias range can be varied in all the amps from about -5 Volts ( near Full Cathode Bias ) to about -35 Volts or so, depending on the amplifier.

    The BENDER-1v. is similar to the BENDER-3v, but will probably use two separate tubes for the voltage gain stage and the Cathode Follower. When the grid of the SV-811 goes positive, it requires considerable current on large signal positive excursions, so the driver tube was chosen to be able to supply large currents, to avoid either frequency response anomalies or current starving in the drive. Note this is still in the early design stage.

    The cathode resistor on the Driver is sized to pull about 20mA DC thru the cathode follower, to exceed the expected worst case peak grid current draw in the SV811-10 S grid current in the SV-811 circuit, for the nominal Bias point at 0VDC.

    The Bias Circuit is not used in the BENDER-1v as the SV-811's grid runs around Zero Volts.

    Biasing The Output Stages

    On the Single Ended designs, the amplifier runs Class A, and the Bias Level would be set to approximately 90% of the tube's rated Plate Dissipation.

    On the Push-Pull design, the operating parameters allow for maximum power, with a minimal crossover distortion by biasing both tubes into conduction near the zero crossover transition point, this is a "high" Class A design.

    The Output Transformers

    The "One-Electron UBT-1" is used on the BENDER-3v., and the "One-Electron UBT-2" or Hammond 1628SE on the BENDER-1v. These have a Primary Impedance of around 1.5K and 5.0K respectively. The

    BENDER-2pp., can be designed using most any "junker" output transformer from an amp or receiver that was rated from 20 Watts to 35 Watts per channel. I used the trannys from an older Pioneer receiver that had been otherwise tossed. It had used pairs of 7868 Novar tubes.

    Reasonable store bought iron for this amp would likely be a 20 Watt Hammond Model 1620 transformer, or similar transformer that physically fits in the chassis space available.

    In each case, the loading gives close to the maximum power output. There is no right or wrong load lines here, just lower distortion or higher output. Lower distortion is obtained at the at the expense of less power output.

    The different taps can be used to make a number of loading possibilities, your speakers are perhaps the final determinants. Experimentation is encouraged here.

    Power Supplies

    The High Voltage power supply section in each of these amplifiers is similar. A pair of Fast Recovery silicon diodes is used in a Voltage Doubler configuration, followed by a Pi network of "C", then "R", then more "C" ( The Cap Bank ).

    Ripple having been reduced, the use of high capacitance, low physical sized Photo-Flash Caps ( only in the later stage of the PS ) in the CAP Bank for each channel in the amp, uses the best possible configuration. This is a simple, but effective arrangement ( which physically fits and works great! )...

    The Power Supply is designed using the Dynaco's conservatively rated power transformer, which has been rewired to operate in isolation mode.

    Fast Recovery silicon diodes provide high charging currents, which would be unavailable using tube rectifiers. The resulting Very Low Impedance Power Supply, looking back, is a much more solid combination of impedances than would be found in almost any low power Tube Amplifier.

    The design uses high current wiring and grounding, as its power supply, in analogous form, is like a huge resevoir that the circuit can call upon for power.

    With a total of about 2,000 uF. of capacitance in the high voltage supply, this is typically about 30 times as much as is found in popular tube amplifiers of the 1960's and 70's.

    The BENDER-3v. and BENDER-2pp. tube filaments are powered from DC Switching Power Supplies.

    This provides a D.C. voltage that remains constant, even as the A.C. line is varied up and down, over a range of beyond 2:1, from below 65 Volts to above 130 Volts for people on the USA side of the pond.

    On the other hand, in the planned BENDER-1v. amplifier, the extremely high filament currents required by the SV811-10's requires a filament supply that can provide 6.3 Volts at 10 Amperes. As no compact Switching Supply of that voltage and ampere capability could be found, this unit uses a conventional brute-force linear D.C. Supply.

    A 16 Ampere power transformer, a chassis mounted 35 Amp bridge rectifier, and 50,000 uF. of multiple high capacitance ( and low ESR ) capacitors do an adequate job of D.C. powering the filaments. Thus both the Directly Heated Triodes and the Input Tubes filaments are DC powered to avoid hum.

    Construction Details

    The entire amp resides in the Steel Chassis and Cage, with the exception of the tips of the glass bottles of the input and output tubes, which stick out of the cage in all the designs.

    Detailed photos are in the articles, more in the Extended Articles and Construction Manuals for each of the above amplifiers, some more photos are on the Web Pages above.

    Actual Performance and More Info

    For more info, you'll either have to read the Construction Articles, or send for a copy of the Extended Articles < $25. > [ when available ] or get the Construction Manuals or Video Tapes [ when available ].

    This page is under construction ( Like software, the job is never done!....)
    Prior Update - 04-01-2001 6:00 PM.
    Current Revision - November 2nd, 2010 @ 10:00 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