I have dealt with the various problems of this design approach many times on AK, either with those looking to have the power amplifier section of interconnected console pieces operate in stand alone fashion, or with those whose output tubes are (literally) burning up in such designs. An example of an in depth project that takes an amplifier designed around this concept and modifies it for maximum performance can be found here:
https://www.audiokarma.org/forums/index.php?threads/improving-the-fisher-x-101c.582379/
But dealing with the problems of this type of design in the manner discussed in that thread is often more than many AKers want to take on or feel comfortable implementing.
With renewed interest being shown in the problems of this biasing concept again, I thought it would be an opportune time to introduce a much simpler modification I devised to alleviate the problems of these designs, so that output tube life and operating temps would no longer be something you'd cringe about every time you operate the amplifier. The modification won't improve performance over the original design as was achieved in the thread above, but original performance levels will be maintained, and your tubes and transformers will breathe a huge sigh of relief just the same. But first, an understanding of just what the problem is.
The concept seems so simple: Use the output tube quiescent current to power the heaters in the small signal tubes. What a win-win deal! DC current is applied to the heaters of the small signal tubes to eliminate any possibility of hum from operating them on AC, and it doesn't require an additional power supply to do it. How cool it that? It's so simple, what could go wrong? Well, that's coming up. Before that however, a few basics.
Fisher used this scheme in numerous stand alone and console models with both 7591 and 6BQ5 class tubes. With high transconductance and similar biasing requirements, both of these tube classes -- when used in push-pull stereo designs -- have collective quiescent current requirements that are well in the ball park of that needed to light a 12AX7 heater when operated in its 12 volt configuration. By adding just a little more voltage to the power transformer's high voltage winding then, it was easy enough to apply this scheme to two, three, and even four series connected 12AX7 heaters to obtain the benefits of DC operated heaters.
The only apparent down side was warmup time: The output tubes had to warm up before they could pass any current to then allow the small signal tubes to warm up. In models that employed a 5AR4 rectifier tube, the warmup time was extended even further because this tube intentionally takes longer to warm up than typical output tubes do. And, because Fisher always operated their DC powered small signal tube heaters at less than their nominally rated voltage (typically in the 10-11 vdc range to maximize signal to noise ratio), this just added insult to injury in the warmup time department. As a boy, we'd turn on Dad's X-101ST (which included all these factors) on Friday, if we wanted to listen to records on Saturday. An exaggeration of course -- but not by much!
Ultimately, it's easy enough to get beyond the warmup time issue, but there was one other pesky little problem to deal with -- and that problem is ultimately what causes so much angst about how the output tubes operate in these models today -- and the problem is not simply higher AC line voltages, although that surely aggravates the issue.
All Fisher amplifiers operate the output stage in Class AB mode, a class characterized by high efficiency and low distortion. The down side of Class AB operation however is that as power output increases, the current drawn by the output stage increases accordingly -- and significantly. Therefore, a 6BQ5 design that might otherwise have a collective cathode current of 130-135 mA under quiescent conditions, could have a collective cathode current of 250 mA with both channels operating at full power. Oops. 12AX7 heaters won't last long in that scenario! Clearly, as the output stages draw more current, the voltage drop across the tube heaters goes up (since they are a resistive element) -- endangering the tubes, and reducing power output in the process.
To guard against this situation from getting out of hand, Fisher always strapped a high wattage resistor across the cathode heater string. This resistor serves one purpose: it prevents the cathode voltage from rising significantly as power output is increased in both channels. This safeguards the small signal tube heaters and, prevents the loss of power output from the increasing voltage drop that would otherwise occur across the heaters in this scenario without the resistor installed. But there's just one problem. The resistor effectively amounts to a shunt regulator, which is never very efficient.
Inefficiency means heat in any electronic device. In my own subject X-101D that I developed this modification on, the stock shunt regulator scheme draws an incredible 37.3 watts of power from the power supply, and then dissipates that energy in (primarily) the output tubes, and the shunt resistor as well -- 33.9 watts in the tubes, and 3.4 watts in the resistor. On a per tube basis, that means that each output tube is dissipating 8.5 watts of energy that's doing nothing but protecting the small signal tube heaters, and maintaining a few extra watts of audio power. Then of course, there's the added heat under the chassis from the resistor, and also the heat not accounted for yet that's developed in the power transformer in working to supply all this extra power. The result is that even with the unit operating from 115 vac, the well balanced output tubes are each dissipating 19.55 watts, or 103% of their Design Maximum Pd rating in the stock design. No wonder the most common complaint today about this design scheme is red platting output tubes!
The goal then is to develop a way to protect the small signal tube heaters and maintain the power output of which the unit is rated for (at least as well as the stock design did), and all without over-dissipating the output tubes, adding any under chassis heat, or cooking the power transformer in the process. The modification proposed here does just that, and is easy to implement. And, with a slight alteration to the parts used, the modification works in either 6BQ5 or 7591 based designs, regardless of how many heaters are (or are not) strung together in the output tube cathode circuit. The remedy to all these problems is the Zener Diode, which is ready made for this application --well that, and the characteristics of a 12AX7 heater. Next time.
Dave
The volunteer test unit the modification was developed on:
https://www.audiokarma.org/forums/index.php?threads/improving-the-fisher-x-101c.582379/
But dealing with the problems of this type of design in the manner discussed in that thread is often more than many AKers want to take on or feel comfortable implementing.
With renewed interest being shown in the problems of this biasing concept again, I thought it would be an opportune time to introduce a much simpler modification I devised to alleviate the problems of these designs, so that output tube life and operating temps would no longer be something you'd cringe about every time you operate the amplifier. The modification won't improve performance over the original design as was achieved in the thread above, but original performance levels will be maintained, and your tubes and transformers will breathe a huge sigh of relief just the same. But first, an understanding of just what the problem is.
The concept seems so simple: Use the output tube quiescent current to power the heaters in the small signal tubes. What a win-win deal! DC current is applied to the heaters of the small signal tubes to eliminate any possibility of hum from operating them on AC, and it doesn't require an additional power supply to do it. How cool it that? It's so simple, what could go wrong? Well, that's coming up. Before that however, a few basics.
Fisher used this scheme in numerous stand alone and console models with both 7591 and 6BQ5 class tubes. With high transconductance and similar biasing requirements, both of these tube classes -- when used in push-pull stereo designs -- have collective quiescent current requirements that are well in the ball park of that needed to light a 12AX7 heater when operated in its 12 volt configuration. By adding just a little more voltage to the power transformer's high voltage winding then, it was easy enough to apply this scheme to two, three, and even four series connected 12AX7 heaters to obtain the benefits of DC operated heaters.
The only apparent down side was warmup time: The output tubes had to warm up before they could pass any current to then allow the small signal tubes to warm up. In models that employed a 5AR4 rectifier tube, the warmup time was extended even further because this tube intentionally takes longer to warm up than typical output tubes do. And, because Fisher always operated their DC powered small signal tube heaters at less than their nominally rated voltage (typically in the 10-11 vdc range to maximize signal to noise ratio), this just added insult to injury in the warmup time department. As a boy, we'd turn on Dad's X-101ST (which included all these factors) on Friday, if we wanted to listen to records on Saturday. An exaggeration of course -- but not by much!
Ultimately, it's easy enough to get beyond the warmup time issue, but there was one other pesky little problem to deal with -- and that problem is ultimately what causes so much angst about how the output tubes operate in these models today -- and the problem is not simply higher AC line voltages, although that surely aggravates the issue.
All Fisher amplifiers operate the output stage in Class AB mode, a class characterized by high efficiency and low distortion. The down side of Class AB operation however is that as power output increases, the current drawn by the output stage increases accordingly -- and significantly. Therefore, a 6BQ5 design that might otherwise have a collective cathode current of 130-135 mA under quiescent conditions, could have a collective cathode current of 250 mA with both channels operating at full power. Oops. 12AX7 heaters won't last long in that scenario! Clearly, as the output stages draw more current, the voltage drop across the tube heaters goes up (since they are a resistive element) -- endangering the tubes, and reducing power output in the process.
To guard against this situation from getting out of hand, Fisher always strapped a high wattage resistor across the cathode heater string. This resistor serves one purpose: it prevents the cathode voltage from rising significantly as power output is increased in both channels. This safeguards the small signal tube heaters and, prevents the loss of power output from the increasing voltage drop that would otherwise occur across the heaters in this scenario without the resistor installed. But there's just one problem. The resistor effectively amounts to a shunt regulator, which is never very efficient.
Inefficiency means heat in any electronic device. In my own subject X-101D that I developed this modification on, the stock shunt regulator scheme draws an incredible 37.3 watts of power from the power supply, and then dissipates that energy in (primarily) the output tubes, and the shunt resistor as well -- 33.9 watts in the tubes, and 3.4 watts in the resistor. On a per tube basis, that means that each output tube is dissipating 8.5 watts of energy that's doing nothing but protecting the small signal tube heaters, and maintaining a few extra watts of audio power. Then of course, there's the added heat under the chassis from the resistor, and also the heat not accounted for yet that's developed in the power transformer in working to supply all this extra power. The result is that even with the unit operating from 115 vac, the well balanced output tubes are each dissipating 19.55 watts, or 103% of their Design Maximum Pd rating in the stock design. No wonder the most common complaint today about this design scheme is red platting output tubes!
The goal then is to develop a way to protect the small signal tube heaters and maintain the power output of which the unit is rated for (at least as well as the stock design did), and all without over-dissipating the output tubes, adding any under chassis heat, or cooking the power transformer in the process. The modification proposed here does just that, and is easy to implement. And, with a slight alteration to the parts used, the modification works in either 6BQ5 or 7591 based designs, regardless of how many heaters are (or are not) strung together in the output tube cathode circuit. The remedy to all these problems is the Zener Diode, which is ready made for this application --well that, and the characteristics of a 12AX7 heater. Next time.
Dave
The volunteer test unit the modification was developed on:
. With all this valuable info, I think you have enough history to make a book where is sure I'll buy one copy.
