Oh Jay -- That'll never do!
Frannie -- vintage electrons come from vintage power stations -- everybody knows that!
Dave
Frannie -- vintage electrons come from vintage power stations -- everybody knows that!
Dave
Teaser Pics -- Fisher SA-100 clone in progress
- Thread starter dcgillespie
- Start date
Progress Update
I have had some more time to devote to this project, so it is moving along well now.
Besides all the mechanical work being completed, now all of the wiring is complete -- that literally being all the wire leads to be installed, which is significant. It represents at least 50% of the total wiring process, and must be done first if any kind of a neat finished job is to achieved.
Also installed are the two pieces of sand comprising the EFB II components for pentode operation, and a small bridge rectifier and regulator to facilitate the conventional dc bias and balance controls for each channel.
From here, installation of all the various Rs and Cs will commence, followed by initial testing to develop the NFB and stability circuits. Once that is done, the final wiring can be completed, and performance fully documented. Pics of existing progress provided:
1. The obligatory glamor shot. But this time with the real tubes that will be used with this project.
2. The rear end of business, just so you know it has a rear end.
3. The power supply end, showing the AC balance, DC bias, EFB range, and Screen voltage controls. Also shown is the Mosfet element of EFB II for control of the screens. Since this build is also serving as part development mule, the EFB range and Screen voltage controls have been included to help determine optimum settings for these parameters. Once determined, they would not be needed for future scratch builds.
4. The power output end of business, showing the DC balance controls and test point terminals. Just like with a kit, some connections are soldered as finished connections, while others are not, awaiting installation of the remaining components.
5. The single cathode EFB(tm) regulator is mounted on the rear panel. Even though this version of EFB has independent DC bias and balance controls for each channel, only one regulator is required to facilitate these controls and EFB action in both channels.
I hope to be reporting back in a couple of weeks on the finished project's documented performance and sonic impressions. I'm having a lot of fun with this project, so I'm taking my time and enjoying it along the way.
Dave
I have had some more time to devote to this project, so it is moving along well now.
Besides all the mechanical work being completed, now all of the wiring is complete -- that literally being all the wire leads to be installed, which is significant. It represents at least 50% of the total wiring process, and must be done first if any kind of a neat finished job is to achieved.
Also installed are the two pieces of sand comprising the EFB II components for pentode operation, and a small bridge rectifier and regulator to facilitate the conventional dc bias and balance controls for each channel.
From here, installation of all the various Rs and Cs will commence, followed by initial testing to develop the NFB and stability circuits. Once that is done, the final wiring can be completed, and performance fully documented. Pics of existing progress provided:
1. The obligatory glamor shot. But this time with the real tubes that will be used with this project.
2. The rear end of business, just so you know it has a rear end.
3. The power supply end, showing the AC balance, DC bias, EFB range, and Screen voltage controls. Also shown is the Mosfet element of EFB II for control of the screens. Since this build is also serving as part development mule, the EFB range and Screen voltage controls have been included to help determine optimum settings for these parameters. Once determined, they would not be needed for future scratch builds.
4. The power output end of business, showing the DC balance controls and test point terminals. Just like with a kit, some connections are soldered as finished connections, while others are not, awaiting installation of the remaining components.
5. The single cathode EFB(tm) regulator is mounted on the rear panel. Even though this version of EFB has independent DC bias and balance controls for each channel, only one regulator is required to facilitate these controls and EFB action in both channels.
I hope to be reporting back in a couple of weeks on the finished project's documented performance and sonic impressions. I'm having a lot of fun with this project, so I'm taking my time and enjoying it along the way.
Dave
Attachments
Lookin' gorgeous, Dave!! :yes:
Hi Frannie -- Yeah, the bench and lab was getting a little crowded, so I had to move the amp off the bench to get a better angle for the underside pics.
The stand behind the amp is my power output tube tester. It will test all traditional power output tubes, yielding brutally honest results. It operates a device under test (DUT) with all elements powered from highly regulated power supplies (including the heater), and conducts the test in Class AB1 pentode mode at real world operating voltages (400 vdc) and dissipation levels (70% of design center ratings), using fixed bias operation. This makes the results completely independent of AC line voltage conditions, as well as DUT construction tolerances (and therefore very repeatable), and also very honest. When the test is conducted, the DUT is driven to the onset of clipping (as viewed on the small scope next to it) by an internal signal generator and driver stage. The dvm on the right is then used with an internal peak detector circuit to measure the peak voltage developed at the plate, from which true power output can then be determined. After 20+ years of measuring all manner of new and used tubes with it, I have developed standards of power output for the various tubes it will test, along with an appropriate English scale (Bad-?-Good as a percent of average new power output) to develop an accurate assessment of existing capability, versus that when new. For my own tubes, I can mark their performance from when they are first placed into service, and then track their performance decline over time against their hours of use.
It is really more of a performance tube tester than a traditional tester. All tubes are first run through my Hickok or B&K testers for a general assessment of gas, shorts, or basic operation, such that by the time a tube is placed in the power output tester, it has already passed the basic health tests from my commercial testers.
Thanks for the interest!
Dave
The stand behind the amp is my power output tube tester. It will test all traditional power output tubes, yielding brutally honest results. It operates a device under test (DUT) with all elements powered from highly regulated power supplies (including the heater), and conducts the test in Class AB1 pentode mode at real world operating voltages (400 vdc) and dissipation levels (70% of design center ratings), using fixed bias operation. This makes the results completely independent of AC line voltage conditions, as well as DUT construction tolerances (and therefore very repeatable), and also very honest. When the test is conducted, the DUT is driven to the onset of clipping (as viewed on the small scope next to it) by an internal signal generator and driver stage. The dvm on the right is then used with an internal peak detector circuit to measure the peak voltage developed at the plate, from which true power output can then be determined. After 20+ years of measuring all manner of new and used tubes with it, I have developed standards of power output for the various tubes it will test, along with an appropriate English scale (Bad-?-Good as a percent of average new power output) to develop an accurate assessment of existing capability, versus that when new. For my own tubes, I can mark their performance from when they are first placed into service, and then track their performance decline over time against their hours of use.
It is really more of a performance tube tester than a traditional tester. All tubes are first run through my Hickok or B&K testers for a general assessment of gas, shorts, or basic operation, such that by the time a tube is placed in the power output tester, it has already passed the basic health tests from my commercial testers.
Thanks for the interest!
Dave
Somehow cats always know when another heater is being built 
OK....you clued me into this......so I gotta make a few comments: Like everyone else, I'm drooling over it! Simply gorgeous in every way!! As to the 'construct', I've got a couple of questions; first, in your opening post you mentioned a 'proper choke' I think? That one looks to be VERY proper!! Can you give a few more details on that beast? I just checked my 100 and I don't see a choke on top.....and unless it's hiding underneath, I'm guessing the addition of that is a big help to the power supply. Dave, I don't know if you know Bobby Callahan....he's a pretty big 'tube dealer' in my area (Bethlehem, MD)....but he also has a 'hot-rodded' 100....and his has some MASSIVE power supply up-grades....but (obviously) no EFB II. Clearly, this already great amp will benefit from your EFBII upgrade, but I'm also wondering about the choke? Also, I'm curious about the specific regulators you're using; are these just generic RS versions....or are you using something of a higher quality level so as to be a bit more confident in their long-term performance, not to mention as a safe-guard against potential 'resultant damage' if failure of these were to occur? I'm completely 'eatin' the dog food' on this EFB upgrade......and see it as a big step forward in the idea of 'modernizing' vintage equipment to make old designs perform even better. But, I'm also somewhat skeptical of the quality level of a lot of the 'silicon' based components that are so mass-produced these days. Anyway.......simply a beautiful job on this amp Dave. I'd say you've out-done yourself....but I suspect if you posted some pics of some of your other builds, including your 'monster' amps.....this just becomes another 'Rembrandt', Van Gogh, or DCPicasso. Best......... Tom D.
Thanks to all for the compliments! There are so many here who do such wonderful work that I have no doubt that it keeps us all inspired.
Tom -- by a "proper choke", I simply meant that the one I am using is fully capable of maintaining its inductance rating even with both channels operating at full power output. So many times, you will see a much smaller, under-rated choke for the job. The stock Dynaco ST70 is a good example of this. That unit uses a 300 ma choke which certainly helps to quiet PS noise at low or no volume and "normal" listening levels. But if the amplifier is driven to full power output in both channels, the choke is overloaded and saturates. Granted, any increase in PS noise is covered by the sound level produced, but the effect is still plainly visible on a scope, where driving both channels to full power output will cause some noise to appear on sine waves since the choke is no longer providing any effective filtering at that point. The choke I am using on this project is not over sized by specification, but simply correctly sized for the task at hand.
As you note, the original SA-100 does not use a choke, no doubt as a cost, weight, and space saving measure. As a general rule, Fisher did not use PS chokes in their audio equipment. However, since I am building my own clone version, I am taking some liberties to improve on the original design where possible, and since plenty of B+ voltage is available, it got added to the design.
As to the regulators, they are very much garden variety devices. I am one who is not opposed to using SS devices with vacuum tube electronics when it makes good and perfect sense to do so. I don't use them indiscriminately, but where good engineering practice would dictate. Even then however, I am very cautious to so over rate the devices so as to virtually never given them a second thought from a reliability standpoint. In this application the 337 EFB(tm) regulator dissipates ~ .5 watt under quiescent conditions, and only 3.9 watts at full sustained power output with both channels operating. This is a 15 watt device, so under worst case continuous conditions, it is only operating at 26% of its base rating, and uses the coolest part of the chassis to dissipate its heat into. Therefore, even with de-rating, it is still a well over rated device for the job.
The other regulator is a 15 watt 317T device that dissipates a constant .03 watt (not a misprint) under any and all conditions of use. Any time spent discussing the stress on this device could better be spent watching grass grow.
Finally, there is also a power mosfet used in the execution of EFB II. This is a readily available 135 watt 9 Amp device that dissipates 2 watts under quiescent conditions, and ~ 4 watts at full sustained power output with both channels operating. Also using the chassis for a heat sink, it is also very well over rated as well.
All three of these devices work in the push-pull (common mode) connection of the output stage, which makes their sonic presence basically invisible. To that point, after two years of numerous installations, any concern over whether EFB would leave a SS footprint in the original Dynaco SCA-35 application has never materialized to my knowledge. The improvements it provides however have been well documented to date.
Collectively then, these devices dissipate about 2.5 watts under quiescent and normal use conditions, and just under 8 watts at full sustained power output with both channels operating. Compare this then, to the original SA-100, whose bleeder resistor screen regulator design dissipates 13 watts under quiescent conditions alone! Against this then, the use of the SS devices in EFB II make good sense in reducing inefficiencies of the original design, while offering the possibility of significant performance improvements, both at the same time.
The SA-100 is a wonderful design for its day and age. I am only trying to take the best of what it was, and make it even better.
Dave
Tom -- by a "proper choke", I simply meant that the one I am using is fully capable of maintaining its inductance rating even with both channels operating at full power output. So many times, you will see a much smaller, under-rated choke for the job. The stock Dynaco ST70 is a good example of this. That unit uses a 300 ma choke which certainly helps to quiet PS noise at low or no volume and "normal" listening levels. But if the amplifier is driven to full power output in both channels, the choke is overloaded and saturates. Granted, any increase in PS noise is covered by the sound level produced, but the effect is still plainly visible on a scope, where driving both channels to full power output will cause some noise to appear on sine waves since the choke is no longer providing any effective filtering at that point. The choke I am using on this project is not over sized by specification, but simply correctly sized for the task at hand.
As you note, the original SA-100 does not use a choke, no doubt as a cost, weight, and space saving measure. As a general rule, Fisher did not use PS chokes in their audio equipment. However, since I am building my own clone version, I am taking some liberties to improve on the original design where possible, and since plenty of B+ voltage is available, it got added to the design.
As to the regulators, they are very much garden variety devices. I am one who is not opposed to using SS devices with vacuum tube electronics when it makes good and perfect sense to do so. I don't use them indiscriminately, but where good engineering practice would dictate. Even then however, I am very cautious to so over rate the devices so as to virtually never given them a second thought from a reliability standpoint. In this application the 337 EFB(tm) regulator dissipates ~ .5 watt under quiescent conditions, and only 3.9 watts at full sustained power output with both channels operating. This is a 15 watt device, so under worst case continuous conditions, it is only operating at 26% of its base rating, and uses the coolest part of the chassis to dissipate its heat into. Therefore, even with de-rating, it is still a well over rated device for the job.
The other regulator is a 15 watt 317T device that dissipates a constant .03 watt (not a misprint) under any and all conditions of use. Any time spent discussing the stress on this device could better be spent watching grass grow.
Finally, there is also a power mosfet used in the execution of EFB II. This is a readily available 135 watt 9 Amp device that dissipates 2 watts under quiescent conditions, and ~ 4 watts at full sustained power output with both channels operating. Also using the chassis for a heat sink, it is also very well over rated as well.
All three of these devices work in the push-pull (common mode) connection of the output stage, which makes their sonic presence basically invisible. To that point, after two years of numerous installations, any concern over whether EFB would leave a SS footprint in the original Dynaco SCA-35 application has never materialized to my knowledge. The improvements it provides however have been well documented to date.
Collectively then, these devices dissipate about 2.5 watts under quiescent and normal use conditions, and just under 8 watts at full sustained power output with both channels operating. Compare this then, to the original SA-100, whose bleeder resistor screen regulator design dissipates 13 watts under quiescent conditions alone! Against this then, the use of the SS devices in EFB II make good sense in reducing inefficiencies of the original design, while offering the possibility of significant performance improvements, both at the same time.
The SA-100 is a wonderful design for its day and age. I am only trying to take the best of what it was, and make it even better.
Dave
Dave.....simply amazing insight! Those tube 'purists' should note that the SS devices are NOT in the signal path.....so in reality.....they're only doing what was either done with other passive parts previously......or, what 'should' have been done back in the day, but wasn't due to either lack of technology, lack of parts, economy of production, or any combination there-of. BTW, you've got me thinkin' about 'chokes' now! Perhaps in the future you can give me some suggestions for my ST-70s. OK...gotta go pack for my trip to Detroit! Early start tomorrow, business stop along the way....and about 10 hours of drive time altogether. Fun!! Later,
Tom D.
Tom D.
wa2ise
Super Member
I am sure that Banditt will guard the Scott amplifier that is belly up next to him
Cats like audio equipment.
Take a look at this AK thread http://www.audiokarma.org/forums/showthread.php?t=193866&highlight=cats+audio and maybe you'd like to add your cat to the thread. :thmbsp:
Boy do they ever! I wasn't aware of that thread, but certainly enjoyed spotting some the pics. Thanks for alerting me to it. That's got to be one of the longest threads on AK. Rather obvious that music, audio equipment, and pets all seem to go hand in hand!
Dave
Dave
More Progress
At this point, the power supply is completely finished except for one resistor feeding B+ to the AF amplifier stage. It has been roughly determined and will be tacked in during the testing phase, with final installation performed once the optimum value has been determined.
Finished also are the EFB II circuits and wiring of the output stages proper.
All that remains then are the phase inverter and AF amplifier circuits, along with the NFB and stability circuits. I will likely only finish one channel from here on out to develop the latter two circuits with, and to determine optimum settings for screen voltage and quiescent current. Once those values/settings are determined, then the other channel can be finished out rather quickly, and final testing performed.
Pics include:
1. Power supply wiring. Metal film resistors used primarily to enhance stability in the EFB II circuits. It is plenty stable without them, so they are icing on the cake.
2. Power output stage wiring. Carbon film does well in the audio end of business, but the cathode resistors are metal film for stability, and mounted away from the tube sockets to further enhance stability as well.
3. An overall shot of chassis wiring to date. The underside is filling up, but hardly crammed considering circuit features and the use of point-to-point wiring.
4. Time for more testing to verify the new wiring. Garden variety (i.e lab rat duty) Russian 6P14Ps are installed for this purpose. The basic AC, heater, rectifier tube, and B+ wiring (up through the choke) have already been previously checked out.
5. Basically a rather boring start up. The B+ ramped up uneventfully, and each tube could easily be set for an initial 25 ma quiescent current draw. Basic EFB II action was verified, and the unit left to cook for a couple of hours. After that time, the chassis area around the EFB regulator and mosfet device were only mildly warm to the touch, the tubes were all still pulling 25 ma quiescent current, and the power transformer could only be considered warm at best.
With all the installed circuits verified, I'm well on my way to third now. The expectations are high for the final product, after which I hope to be able to pass on the resulting mods for SA-100 owners, which should offer real performance improvements for these units. Those will be outlined in the followup article to be written once the amplifier project is complete.
Dave
At this point, the power supply is completely finished except for one resistor feeding B+ to the AF amplifier stage. It has been roughly determined and will be tacked in during the testing phase, with final installation performed once the optimum value has been determined.
Finished also are the EFB II circuits and wiring of the output stages proper.
All that remains then are the phase inverter and AF amplifier circuits, along with the NFB and stability circuits. I will likely only finish one channel from here on out to develop the latter two circuits with, and to determine optimum settings for screen voltage and quiescent current. Once those values/settings are determined, then the other channel can be finished out rather quickly, and final testing performed.
Pics include:
1. Power supply wiring. Metal film resistors used primarily to enhance stability in the EFB II circuits. It is plenty stable without them, so they are icing on the cake.
2. Power output stage wiring. Carbon film does well in the audio end of business, but the cathode resistors are metal film for stability, and mounted away from the tube sockets to further enhance stability as well.
3. An overall shot of chassis wiring to date. The underside is filling up, but hardly crammed considering circuit features and the use of point-to-point wiring.
4. Time for more testing to verify the new wiring. Garden variety (i.e lab rat duty) Russian 6P14Ps are installed for this purpose. The basic AC, heater, rectifier tube, and B+ wiring (up through the choke) have already been previously checked out.
5. Basically a rather boring start up. The B+ ramped up uneventfully, and each tube could easily be set for an initial 25 ma quiescent current draw. Basic EFB II action was verified, and the unit left to cook for a couple of hours. After that time, the chassis area around the EFB regulator and mosfet device were only mildly warm to the touch, the tubes were all still pulling 25 ma quiescent current, and the power transformer could only be considered warm at best.
With all the installed circuits verified, I'm well on my way to third now. The expectations are high for the final product, after which I hope to be able to pass on the resulting mods for SA-100 owners, which should offer real performance improvements for these units. Those will be outlined in the followup article to be written once the amplifier project is complete.
Dave
Attachments
Finished!
My Fisher SA-100 clone with EFB II is all finished now. The final steps of finishing out one basic channel, developing and installing the NFB and stability circuits, and then finishing out the other channel are all complete -- as is final testing and documentation as well.
To recap, this project follows the basic topology of the Fisher SA-100 design (GZ34 tube rectification, push-pull 7189 fixed bias pentode operated output stages, 7247 driver tubes for AF amplifier and cathodyne phase inverter, and ~ 20 db global NFB), but uses various parts and pieces from other equipment in the build. The power transformer came from a Fisher X-101 ST, and the OPTs are from a Heath AA-100 7591 integrated amplifier. These OPTs have a lam stack that is identical in size to the transformers used in the SA-100, and have similar impedance characteristics as well. Additional features in my project include the use of a choke that can handle all the current needs of the amplifier, splitting the original 7247 driver tubes into a single 12AX7 and 6FQ7 for practicality, and the implementation of EFB II for pentode based fixed bias output stages. Like the original EFB(tm) concept for UL applications, EFB II not only locks in the output tube bias/B+ relationship regardless of the actual B+ level present, but also maintains the screen/B+ relationship as well -- all in addition to providing fixed bias operation. This means that the lowest distortion operating conditions for the output tubes are maintained regardless of variations in the B+ supply -- be it due to changing AC line conditions, or due to changing dynamic conditions (power output) in the amplifier itself. It is all built on a Hammond chassis. The measured performance is quite high, and speaks for itself:
1. Maximum power output is 20.0 watts RMS per channel, with both channels continuously driven, over a frequency range of 25 Hz to 20 kHz. Under the same test conditions, each channel produces just over 18 watts RMS at 20 Hz.
(Individually driven channel measurements indicate that 22 watts RMS is developed over the same power bandwidth, with no power increase at 20 Hz due to OPT limitations)
2. Total Harmonic Distortion over the 25 Hz to 20 kHz power bandwidth is less than 1.0% within 1 db of 20 watts RMS power output (about 16 watts RMS per channel), with both channels driven.
At 20 Hz, 10 watts RMS is developed per channel with no more than 1% THD, with both channels driven.
3. The frequency response of both channels is +0/-1 db to 70 kHz, measured at a standard 1 watt power output level.
4. Stability is such that no amount of capacitive loading (only) on the output will cause any tendency towards oscillation.
5. 10 kHz square waves display a fast rise time on the leading edge, and well damped wave top. A couple pics are provided below.
6. The clipping characteristics are smooth and even, and recovery from overload is virtually instantaneous.
This is very high performance indeed. After listening to it for about 11 hours now, it displays all the dynamics and sharp detail that the specifications would imply in an effortless fashion, without any tendency towards stridency or muddled bass. Definition is very high over the entire frequency spectrum (that I can hear anyway) through my Cornwalls. It simply provides all the sound so loved from the EL84 family of tubes, but gives every impression of presenting it with the authority of the best 7591 designs.
For my article, I am working with a couple of fellow AKers to include performance data in it from not only my own build, but also from an authentic stock Fisher SA-100, and from a different (but superb) diy build of the original SA-100 design -- again using Heath AA-100 OPTs, but using a very stiff SS power supply instead, and no EFB. This should not only allow comparison of the two diy builds to the original Fisher product, but also of a conventional tube rectified power supply (such as the SA-100 uses) with EFB, versus that of a stiff SS power supply without EFB.
If anything, it should hopefully make for interesting reading, but also describe EFB II in detail, and show how it can easily be applied to literally gazillions of push-pull EL84 amplifiers out there for a nice performance improvement.
The pics provided below include:
1. On the test bench during development of the NFB and stability circuits. Gotta love alligator clip leads! Numerous components are still just tacked into place.
2. The finished build. The testing is all done, and all connections are soldered. Since this was originally a junk box build, some components will likely be changed out later on for higher quality parts (same values), but for now, it's time to enjoy.
3. This is a 10 kHz square wave (left channel in this case, but the right channel is identical) as produced across a resistively loaded output. This is a very good, classic, square wave presentation at this frequency, and indicates many things about the design's frequency response and transient performance. The nearly vertical rise followed by the small hump at the top of the wave is typical of high quality vacuum tube NFB amplifiers, indicating a well extended HF response beyond the traditional audio spectrum, necessary for the application of stable NFB. The wave top is very smooth and straight (without wiggles and bumps), indicating the amplifier's ability to damp itself very quickly after a transient spike such as a square wave represents. Were there numerous wiggles and bumps, it indicates that the design is only marginally stable, with peaks and valleys in it's response curve. These are just some of the reasons that square wave testing is so important (music itself is quite transient as well), and can lay bare so many performance characteristics (good and bad!) of a given design.
4. This is a very rigorous test, and one that many, many of your favorite amplifiers will fail miserably at. This is a 10 kHz waveform, as presented at the output of the amplifier (right channel in this case) -- with no other load than a .25 uF cap connected across the output. It is an important test because long speaker lines, big speaker lines, and complicated speaker crossover networks can represent a significant amount of capacitance placed across the output of the amplifier. I have always maintained that how well a NFB vacuum tube amplifier maintains its composure with difficult loads such as this test represents, will have a greater bearing on the resulting sound produced than any amount of tube rolling, cable/component swapping, or anything else that can be changed out will have. Without a normal resistive load on the output, the amplifier's output spikes as it works to charge the cap on each cycle. But notice that even with so rigorous a test as this, the amplifier maintains it's composure, and dampens out the wave well before the end of the wave form. Many, many amplifiers as originally designed will simply burst into sustained supersonic oscillation with such a test, annoying your neighbor's radio reception, and have you quickly applying for an Amateur Radio's license!
5. Up and running in the rotation. This is a wonderful amplifier to listen to. Even after a full day of operation, the transformers are only warm to the touch, and the amplifier never draws attention to itself. It simply does what it is supposed to do very well. With no signal running through it, it is dead quiet right at the speaker board of the Cornwalls. With a signal applied however, it can make its presence known quite quickly!
I have enjoyed this project immensely. Thanks to all of you who have followed along. A break before the real SA-100 arrives, where then it will be on to testing and evaluating the Fisher's actual stock performance. After that, the article can get underway, and will be published on the Tronola website sometime later this year. Until then, enjoy your projects. Good luck with them!
Dave
My Fisher SA-100 clone with EFB II is all finished now. The final steps of finishing out one basic channel, developing and installing the NFB and stability circuits, and then finishing out the other channel are all complete -- as is final testing and documentation as well.
To recap, this project follows the basic topology of the Fisher SA-100 design (GZ34 tube rectification, push-pull 7189 fixed bias pentode operated output stages, 7247 driver tubes for AF amplifier and cathodyne phase inverter, and ~ 20 db global NFB), but uses various parts and pieces from other equipment in the build. The power transformer came from a Fisher X-101 ST, and the OPTs are from a Heath AA-100 7591 integrated amplifier. These OPTs have a lam stack that is identical in size to the transformers used in the SA-100, and have similar impedance characteristics as well. Additional features in my project include the use of a choke that can handle all the current needs of the amplifier, splitting the original 7247 driver tubes into a single 12AX7 and 6FQ7 for practicality, and the implementation of EFB II for pentode based fixed bias output stages. Like the original EFB(tm) concept for UL applications, EFB II not only locks in the output tube bias/B+ relationship regardless of the actual B+ level present, but also maintains the screen/B+ relationship as well -- all in addition to providing fixed bias operation. This means that the lowest distortion operating conditions for the output tubes are maintained regardless of variations in the B+ supply -- be it due to changing AC line conditions, or due to changing dynamic conditions (power output) in the amplifier itself. It is all built on a Hammond chassis. The measured performance is quite high, and speaks for itself:
1. Maximum power output is 20.0 watts RMS per channel, with both channels continuously driven, over a frequency range of 25 Hz to 20 kHz. Under the same test conditions, each channel produces just over 18 watts RMS at 20 Hz.
(Individually driven channel measurements indicate that 22 watts RMS is developed over the same power bandwidth, with no power increase at 20 Hz due to OPT limitations)
2. Total Harmonic Distortion over the 25 Hz to 20 kHz power bandwidth is less than 1.0% within 1 db of 20 watts RMS power output (about 16 watts RMS per channel), with both channels driven.
At 20 Hz, 10 watts RMS is developed per channel with no more than 1% THD, with both channels driven.
3. The frequency response of both channels is +0/-1 db to 70 kHz, measured at a standard 1 watt power output level.
4. Stability is such that no amount of capacitive loading (only) on the output will cause any tendency towards oscillation.
5. 10 kHz square waves display a fast rise time on the leading edge, and well damped wave top. A couple pics are provided below.
6. The clipping characteristics are smooth and even, and recovery from overload is virtually instantaneous.
This is very high performance indeed. After listening to it for about 11 hours now, it displays all the dynamics and sharp detail that the specifications would imply in an effortless fashion, without any tendency towards stridency or muddled bass. Definition is very high over the entire frequency spectrum (that I can hear anyway) through my Cornwalls. It simply provides all the sound so loved from the EL84 family of tubes, but gives every impression of presenting it with the authority of the best 7591 designs.
For my article, I am working with a couple of fellow AKers to include performance data in it from not only my own build, but also from an authentic stock Fisher SA-100, and from a different (but superb) diy build of the original SA-100 design -- again using Heath AA-100 OPTs, but using a very stiff SS power supply instead, and no EFB. This should not only allow comparison of the two diy builds to the original Fisher product, but also of a conventional tube rectified power supply (such as the SA-100 uses) with EFB, versus that of a stiff SS power supply without EFB.
If anything, it should hopefully make for interesting reading, but also describe EFB II in detail, and show how it can easily be applied to literally gazillions of push-pull EL84 amplifiers out there for a nice performance improvement.
The pics provided below include:
1. On the test bench during development of the NFB and stability circuits. Gotta love alligator clip leads! Numerous components are still just tacked into place.
2. The finished build. The testing is all done, and all connections are soldered. Since this was originally a junk box build, some components will likely be changed out later on for higher quality parts (same values), but for now, it's time to enjoy.
3. This is a 10 kHz square wave (left channel in this case, but the right channel is identical) as produced across a resistively loaded output. This is a very good, classic, square wave presentation at this frequency, and indicates many things about the design's frequency response and transient performance. The nearly vertical rise followed by the small hump at the top of the wave is typical of high quality vacuum tube NFB amplifiers, indicating a well extended HF response beyond the traditional audio spectrum, necessary for the application of stable NFB. The wave top is very smooth and straight (without wiggles and bumps), indicating the amplifier's ability to damp itself very quickly after a transient spike such as a square wave represents. Were there numerous wiggles and bumps, it indicates that the design is only marginally stable, with peaks and valleys in it's response curve. These are just some of the reasons that square wave testing is so important (music itself is quite transient as well), and can lay bare so many performance characteristics (good and bad!) of a given design.
4. This is a very rigorous test, and one that many, many of your favorite amplifiers will fail miserably at. This is a 10 kHz waveform, as presented at the output of the amplifier (right channel in this case) -- with no other load than a .25 uF cap connected across the output. It is an important test because long speaker lines, big speaker lines, and complicated speaker crossover networks can represent a significant amount of capacitance placed across the output of the amplifier. I have always maintained that how well a NFB vacuum tube amplifier maintains its composure with difficult loads such as this test represents, will have a greater bearing on the resulting sound produced than any amount of tube rolling, cable/component swapping, or anything else that can be changed out will have. Without a normal resistive load on the output, the amplifier's output spikes as it works to charge the cap on each cycle. But notice that even with so rigorous a test as this, the amplifier maintains it's composure, and dampens out the wave well before the end of the wave form. Many, many amplifiers as originally designed will simply burst into sustained supersonic oscillation with such a test, annoying your neighbor's radio reception, and have you quickly applying for an Amateur Radio's license!
5. Up and running in the rotation. This is a wonderful amplifier to listen to. Even after a full day of operation, the transformers are only warm to the touch, and the amplifier never draws attention to itself. It simply does what it is supposed to do very well. With no signal running through it, it is dead quiet right at the speaker board of the Cornwalls. With a signal applied however, it can make its presence known quite quickly!
I have enjoyed this project immensely. Thanks to all of you who have followed along. A break before the real SA-100 arrives, where then it will be on to testing and evaluating the Fisher's actual stock performance. After that, the article can get underway, and will be published on the Tronola website sometime later this year. Until then, enjoy your projects. Good luck with them!
Dave
