Avery's "First" Maker: The TA-800
- Thread starter dcgillespie
- Start date
From a purely practical standpoint, using a 12AX7 and a 12AU7 makes the most sense. In fact, when I did my SA-100 Clone, I used a 12AX7 and 6CG7/6FQ7 in that very configuration. But that project was a scratch build, whereas this is a bonafide Fisher product. To that point, one of the guiding principles for this part of the the project is a healthy respect for how Fisher would have done this modification if they had in fact done it. We've certainly got the late version 600 to go by -- not only in what tube they changed to, but in how and where they made the changes to accommodate the new tube in the circuit. In discussing all of this with Rob, it was agreed to apply/use as much Fisher design philosophy of that time to the modified circuit as possible, so that the end result still comes off as a "Fisher" -- as seamlessly if Fisher had done it in the first place. In other words, it comes off as a "late version" TA-800, had one ever been produced.
Dave
Dave
Minor request for clarification
Looking at it, it seems like the screen resistor is bypassed with a cap to ground. Its not a very large value though, 0.1uf it appears. Is this where its getting NFB from because the cap isn't large enough to hold the screen at strictly DC? Is there any sort of frequency constant in play here where the value of the cap bypasses more effectively at high frequencies, resulting in more NFB in the bass region? Just asking because I have an amp with a pentode input stage and its a bit overly sensitive. It uses I think a 4uf or so cap from screen to ground. I was thinking if I can reduce the size of that to tickle in a bit of local feedback to drop the gain of the stage it might be useful. I don't want to mess it up in such a way that it has local feedback at the voltage amp stage that kills the bottom end though.
Looking at it, it seems like the screen resistor is bypassed with a cap to ground. Its not a very large value though, 0.1uf it appears. Is this where its getting NFB from because the cap isn't large enough to hold the screen at strictly DC? Is there any sort of frequency constant in play here where the value of the cap bypasses more effectively at high frequencies, resulting in more NFB in the bass region? Just asking because I have an amp with a pentode input stage and its a bit overly sensitive. It uses I think a 4uf or so cap from screen to ground. I was thinking if I can reduce the size of that to tickle in a bit of local feedback to drop the gain of the stage it might be useful. I don't want to mess it up in such a way that it has local feedback at the voltage amp stage that kills the bottom end though.
Hi Gadget -- After I read what I posted, I saw that and thought the more technically minded might question the comment. Glad you asked, as it gives me a chance to explain it further.
NFB is effectively introduced at the screen grid by virtue of the fact that the screen grid itself is held at AC ground potential by way of the .1 uF cap, but the cathode is not, due to the NFB signal inserted there by the NFB loop. It happens as follows:
The NFB signal injected at the cathode is in-phase with the audio signal applied at the control grid, so it reduces (or works to cancel) the effective value of the signal actually applied between the cathode and control grid elements, and therefore reduces the amplitude of the opposite phase signal produced at the plate. The application of NFB then can be seen as a process of cancellation, but the cancellation at the control grid is not the only cancellation taking place.
With the screen grid held at AC ground, it maximizes the amplification of the tube with respect to signals injected at the cathode (i.e., as a grounded grid amplifier), whose signals always appear in phase at the plate. Since the NFB is injected at the cathode, the AC grounded screen grid then maximizes how much the NFB signal is amplified to appear as a component of the plate signal. Since the component of the plate signal that relates to the cathode injected NFB signal is out of phase with the plate signal developed by the net control grid signal, it acts to reduce the amplitude of the plate signal even further -- which any increase in NFB would. In this case, the increase in NFB occurs because holding the screen grid at AC ground level maximizes the effectiveness of the tube as a grounded grid amplifier.
Now. If the screen grid is not held at AC ground level but rather, the screen grid bypass cap is returned to the point where the NFB is injected at the cathode, it changes things considerably. Now, the NFB signal is no longer appearing between the cathode and screen grid as before, so the the NFB applied to the cathode is no longer amplified by the grounded grid configuration of the screen grid, but simply passed on to the plate. Therefore, at the plate, the cathode injected NFB is still countering the signal generated there by the net control grid voltage, but not by nearly as much. Therefore, the plate signal amplitude increases -- by a factor equal to the screen grid Gm of the tube, since it is no longer acting as a grounded (screen) grid amplifier for the cathode injected NFB signal.
From the perspective of the control grid however, when NFB is applied at the cathode input, grounding the screen grid bypass cap causes the the amplification of the tube to be reduced for control grid input signals, because the NFB voltage modulates the screen grid voltage so as to reduce the component of the plate signal produced by the net voltage applied to the control grid. Conversely, when the screen grid bypass cap is attached to the point where the NFB is injected at the cathode however, it maximizes the gain of the tube relative to signals input at the control grid.
When the effect of eliminating the grounded screen grid amplification of the NFB signals is added to maximizing the amplification of control grid signals (both resulting from tying the screen grid bypass cap to the point of NFB insertion), the signal at the plate increases significantly by a factor equal to the increase in control grid Gm, and the screen grid Gm. Even if the FB is increased to counter the loss of grounded grid NFB amplification, the input sensitivity will still increase significantly when the screen grid bypass cap is tied to the point of NFB injection, because the control grid Gm is so much greater than the screen grid Gm.
Ultimately it will be seen then that when the screen grid bypass cap is directly grounded, it results in applying additional feedback due to the tube amplifying the NFB signal presented at the cathode, and reducing the amplification of signals presented to the control grid. For the purposes of this discussion then, since the overall change in the effective gain of the stage is directly the result of how the screen grid is bypassed, the results achieved are effectively the same as if the applicable NFB elements were injected directly into the screen grid circuit itself.
I hope this helps!
Dave
NFB is effectively introduced at the screen grid by virtue of the fact that the screen grid itself is held at AC ground potential by way of the .1 uF cap, but the cathode is not, due to the NFB signal inserted there by the NFB loop. It happens as follows:
The NFB signal injected at the cathode is in-phase with the audio signal applied at the control grid, so it reduces (or works to cancel) the effective value of the signal actually applied between the cathode and control grid elements, and therefore reduces the amplitude of the opposite phase signal produced at the plate. The application of NFB then can be seen as a process of cancellation, but the cancellation at the control grid is not the only cancellation taking place.
With the screen grid held at AC ground, it maximizes the amplification of the tube with respect to signals injected at the cathode (i.e., as a grounded grid amplifier), whose signals always appear in phase at the plate. Since the NFB is injected at the cathode, the AC grounded screen grid then maximizes how much the NFB signal is amplified to appear as a component of the plate signal. Since the component of the plate signal that relates to the cathode injected NFB signal is out of phase with the plate signal developed by the net control grid signal, it acts to reduce the amplitude of the plate signal even further -- which any increase in NFB would. In this case, the increase in NFB occurs because holding the screen grid at AC ground level maximizes the effectiveness of the tube as a grounded grid amplifier.
Now. If the screen grid is not held at AC ground level but rather, the screen grid bypass cap is returned to the point where the NFB is injected at the cathode, it changes things considerably. Now, the NFB signal is no longer appearing between the cathode and screen grid as before, so the the NFB applied to the cathode is no longer amplified by the grounded grid configuration of the screen grid, but simply passed on to the plate. Therefore, at the plate, the cathode injected NFB is still countering the signal generated there by the net control grid voltage, but not by nearly as much. Therefore, the plate signal amplitude increases -- by a factor equal to the screen grid Gm of the tube, since it is no longer acting as a grounded (screen) grid amplifier for the cathode injected NFB signal.
From the perspective of the control grid however, when NFB is applied at the cathode input, grounding the screen grid bypass cap causes the the amplification of the tube to be reduced for control grid input signals, because the NFB voltage modulates the screen grid voltage so as to reduce the component of the plate signal produced by the net voltage applied to the control grid. Conversely, when the screen grid bypass cap is attached to the point where the NFB is injected at the cathode however, it maximizes the gain of the tube relative to signals input at the control grid.
When the effect of eliminating the grounded screen grid amplification of the NFB signals is added to maximizing the amplification of control grid signals (both resulting from tying the screen grid bypass cap to the point of NFB insertion), the signal at the plate increases significantly by a factor equal to the increase in control grid Gm, and the screen grid Gm. Even if the FB is increased to counter the loss of grounded grid NFB amplification, the input sensitivity will still increase significantly when the screen grid bypass cap is tied to the point of NFB injection, because the control grid Gm is so much greater than the screen grid Gm.
Ultimately it will be seen then that when the screen grid bypass cap is directly grounded, it results in applying additional feedback due to the tube amplifying the NFB signal presented at the cathode, and reducing the amplification of signals presented to the control grid. For the purposes of this discussion then, since the overall change in the effective gain of the stage is directly the result of how the screen grid is bypassed, the results achieved are effectively the same as if the applicable NFB elements were injected directly into the screen grid circuit itself.
I hope this helps!
Dave
Last edited:
Humm -- not many of these suckers made.....
Dave
Dave
The old components have been removed from the Channel A driver tube socket now, and the pins cleaned up ready for the new driver circuit to be installed. That hasn't been developed yet since the old circuitry had to be removed before that effort can begin. The only component that remains of the original build is one of the new coupling caps originally installed will remain connected to pin #1. With the old out then, the first order of business is to get things lit up. Therefore, the heater wiring that originally appeared at pins 4&5 has to be reconfigured, and now appears between pins 4&5 tied together, and pin 9 to accommodate the new 7247 tube. This will require the green heater lead coming from the power supply area to be lengthened so as to be able to reach its new destination (pin 9). This lead is the main lead powering all the 6.3 vac tubes, so it is heavier gauge wire than is typically used throughout the unit.
So now development of the new driver stage can begin. The basic topology of the new driver stage is easy enough, and can largely follow that of the later version 600s. It will be modified slightly to accommodate the AC Balance adjustment afforded the TA-800s, that the 600s do not employ. Therefore, that feature will remain. Where the real development work comes in is in establishing the original feedback level with the new driver configuration, and then ensuring it is properly stabilized to produce optimum performance. Fisher pulled a few shenanigans in this regard with both of the TAs, including both versions of the 600 (the later of which employs a different OPT from its earlier version), so development of this portion of the modification starts from scratch.
The pic shows all the old removed, and the heater wiring reconfigured to accommodate the new tube. More as it becomes available.
Dave
So now development of the new driver stage can begin. The basic topology of the new driver stage is easy enough, and can largely follow that of the later version 600s. It will be modified slightly to accommodate the AC Balance adjustment afforded the TA-800s, that the 600s do not employ. Therefore, that feature will remain. Where the real development work comes in is in establishing the original feedback level with the new driver configuration, and then ensuring it is properly stabilized to produce optimum performance. Fisher pulled a few shenanigans in this regard with both of the TAs, including both versions of the 600 (the later of which employs a different OPT from its earlier version), so development of this portion of the modification starts from scratch.
The pic shows all the old removed, and the heater wiring reconfigured to accommodate the new tube. More as it becomes available.
Dave
DAMN! That explanation on POST # 125 had me feeling I was on a circular driveway getting nowhere fast. Gotta read it a few more times to look for an exit.
I was able to pull my TA-800 out far enough to read these numbers. 11739
Looks like there might of been a letter or another number, but it is not clear enough to see exactly what it is.
And just to verify, my buddy Jim has 6BL8's installed with adapters in this receiver.
Looks like there might of been a letter or another number, but it is not clear enough to see exactly what it is.
And just to verify, my buddy Jim has 6BL8's installed with adapters in this receiver.
I do remember his railing about how much 12AU7's suck in the tube forum and how wonderful 6BL8s were and not much else beyond that as I started the "Tune Out"I was able to pull my TA-800 out far enough to read these numbers. 11739
Looks like there might of been a letter or another number, but it is not clear enough to see exactly what it is.
And just to verify, my buddy Jim has 6BL8's installed with adapters in this receiver.
Well Steve at least you are gonna be able to make that right with Dave's expert help here.
Frannie
I was able to get the S/N off of a TA800 that sold on eBay. It was 12186C. I guess mine is the youngest so far at 12944C. A Google image search also found S/N 11575C.
I'm also wondering how many of these were just thrown in a dumpster back in the day because of the lack of a MPX adapter.
I'm also wondering how many of these were just thrown in a dumpster back in the day because of the lack of a MPX adapter.
Last edited:
I can tell you that even though the TA-800 FM is not in stereo, it really sounds great. Since I don't ever equate quality listening with FM, (stations suck here) I have no complaints with either TA series receiver I own as far as FM goes.
Larry -- The explanation is in fact somewhat convoluted, because how the screen grid bypass cap is connected affects the gain of both the control grid input, and the cathode input of the original design-- and does so in opposite directions, for two different signals! So convoluted indeed. I just figured if I was going to cover it, go all the way!
Dave
Dave
Aha, that helps a lot. So basically its because of where the screen cap connects, not the value. I see a lot of things connect the screen cap to the cathode and almost none directly to ground. I never quite knew why this was done. So basically as long as the cap is large enough to provide an effective AC ground across the spectrum, thats all we care about right? Same sort of idea as the grid cap on an LTP inverter stage.
The particular amp I'm fooling with already has it to ground, though in my case it was not so much by original design but by virtue of changes I made. Originally it had no feedback, a grounded cathode, and grid leak bias. I changed it to cathode bias with NFB injected basically like Fisher did it. I did not change the screen cap because the amp uses a circuit board and it would have been very difficult to do so. Good, now I continue to have no reason to do that.
The particular amp I'm fooling with already has it to ground, though in my case it was not so much by original design but by virtue of changes I made. Originally it had no feedback, a grounded cathode, and grid leak bias. I changed it to cathode bias with NFB injected basically like Fisher did it. I did not change the screen cap because the amp uses a circuit board and it would have been very difficult to do so. Good, now I continue to have no reason to do that.
Taking a break from the project this afternoon, but was able to get the basic 7247 driver stage installed now in Channel A. Next up, the process of adding and stabilizing the NFB begins.......
Dave
Dave
Oh, have been meaning to mention the TA-800 Service Manual errors found so far:
1. C140: Shown connected between R157 near V14, and ground. It does not exist, is not installed in the unit, and is not shown in the parts list. X this cap out on your schematic.
2. The Balance control R128 is listed in the parts list, but no value is given. By measurement, it is a dual 300K piece, similar to that used in the TA-600s, but carries a different part number because it is part of the Remote Volume control, which the TA-600 does not have.
3. Similarly, the Remote Volume control (part of the balance control) does not have a value specified with it either. By measurement, it is a 1000Ω control.
4. The bridge diode powering the DC Heater/Bias supply is shown backwards in the schematic.
5. The original installed value at R74 & R84 is 1.8M
Dave
Updated 7-2-2017
1. C140: Shown connected between R157 near V14, and ground. It does not exist, is not installed in the unit, and is not shown in the parts list. X this cap out on your schematic.
2. The Balance control R128 is listed in the parts list, but no value is given. By measurement, it is a dual 300K piece, similar to that used in the TA-600s, but carries a different part number because it is part of the Remote Volume control, which the TA-600 does not have.
3. Similarly, the Remote Volume control (part of the balance control) does not have a value specified with it either. By measurement, it is a 1000Ω control.
4. The bridge diode powering the DC Heater/Bias supply is shown backwards in the schematic.
5. The original installed value at R74 & R84 is 1.8M
Dave
Updated 7-2-2017
Last edited:
fred soop
Super Member
You do realize that Dave composed that while using SIRI, talking into an iPhone while driving around a roundabout.
>> .. ..-. / -.-- --- ..- / -.-. .- -. / .-. . .- -.. / - .... .. ... / -.-- --- ..- / .- .-. . / - --- --- / -.-. .-.. --- ... . <<
-. --- / .. -- / -. --- - / - --- --- / -.-. .-.. --- ... . / -... ..- - / .. / .... .- ...- . / .- / --. --- --- -.. / -- --- -. .. - --- .-.
At the rate you're going, you'll be done before that MPX-100 arrives on Wednesday/Thursday.
FRED: .. ..-. -.-- --- ..- -.-. .- -. .-. . .- -.. - .... .. ... - .... .- -. -.- -.-- --- ..- .-. - . .- -.-. .... . .-. ·-·-·- .. ..-. -.-- --- ..- -.-. .- -. .-. . .- -.. - .... .. ... .. -. . -. --. .-.. .. ... .... - .... .- -. -.- .- -- .- .-. .. -. . -·-·--
FRED: .. ..-. -.-- --- ..- -.-. .- -. .-. . .- -.. - .... .. ... - .... .- -. -.- -.-- --- ..- .-. - . .- -.-. .... . .-. ·-·-·- .. ..-. -.-- --- ..- -.-. .- -. .-. . .- -.. - .... .. ... .. -. . -. --. .-.. .. ... .... - .... .- -. -.- .- -- .- .-. .. -. . -·-·--
Oh heck no. I'm just trying to get back on track to start the EFB work!
Dave
Dave
Development of the new power amplifier driver circuit is all finished now, and fully installed into Channel A. The Channel B driver circuit is still stock, and continues to act as a reference against which to measure the development work in Channel A. Next up, is digging into the Channel A 1st tone control amplifier stage to unlock some gain so that ultimately, the input sensitivity of Channel A will once again equal that of the stock design in Channel B. Once that portion of the modification is developed and installed, the unit then goes back into the listening room -- with Channel A now fully converted over to the new driver tube design, and Channel B still completely stock. With both mono and stereo sources then, the audible performance match between the two channels can be verified, which of course should be identical. Once that is proven out, then all of the modifications developed in Channel A for the new driver tube will be installed in Channel B, to complete this little side venture.
I would also offer that whereas so many of my previous threads and offerings have been singularly focused on (among other things) improving the performance of the original Fisher design, in this case, that took a back seat. Here, the goal was to as accurately as possible, mimic the performance characteristics of the original Fisher driver circuit with a new driver circuit, using a much more practical (for a whole host of reasons) tube to get the job done.
I am always very cognizant that all of these units are first and foremost "Fishers". While I believe that every unit I've modified has always retained its Fisher identity, some no doubt were really pushed in that regard when a number of old (useless) features were removed, new ones installed that were never originally there, and others were re-purposed for more practical use. Other units were modified with purely performance gains in mind, with no other goal sought, although dependability and tube life has always been a part of each and every modification.
For this project then, dependability is a key factor, as Rob apparently wants this unit to be operational into the next century. Adding a little extra kick, promoting cooler operating, and extending tube life are also key elements, which the addition of EFB will produce. But folks who love their Fishers that much love it first for the sound it produces, and are very protective of it. Therefore, open driver stage surgery can easily make them nervous -- wondering if such a significant modification will change the sound of the unit. As a result, as stated above, the goal of this particular portion of the project is to mimic the performance of the original driver circuit (when operating as intended) -- warts and all -- so that the original sonic signature will be maintained. As to the warts of the stock design -- it is somewhat over compensated, which helps to produce that lush Fisher sound that Fisher is famous for. Therefore, the new driver circuit is designed to achieve that same end. The end result came out very well, as the pics provided will show.
More as it unfolds.
Dave
BELOW: These are 10 kHz square waves -- one from the new power amplifier design installed in Channel A, and one from the original stock design in Channel B. They are not perfectly identical, but these displays are more closely matched than many stereo amplifiers that (of course) have the same design in both channels. The new power amplifier design in Channel A is the top display:
BELOW: To show just how closely the two designs are matched however, this shot has one trace from above placed perfectly on top of the other, showing that the rise time, transient response, and flatness of frequency response are all but identical between the two channels, in spite of one channel operating with the old 7199 driver design, and the other with the new 7247 driver:
BELOW: Underside view showing the new driver stage design fully installed in Channel A (closest to the rear of the unit where the level controls are):
I would also offer that whereas so many of my previous threads and offerings have been singularly focused on (among other things) improving the performance of the original Fisher design, in this case, that took a back seat. Here, the goal was to as accurately as possible, mimic the performance characteristics of the original Fisher driver circuit with a new driver circuit, using a much more practical (for a whole host of reasons) tube to get the job done.
I am always very cognizant that all of these units are first and foremost "Fishers". While I believe that every unit I've modified has always retained its Fisher identity, some no doubt were really pushed in that regard when a number of old (useless) features were removed, new ones installed that were never originally there, and others were re-purposed for more practical use. Other units were modified with purely performance gains in mind, with no other goal sought, although dependability and tube life has always been a part of each and every modification.
For this project then, dependability is a key factor, as Rob apparently wants this unit to be operational into the next century. Adding a little extra kick, promoting cooler operating, and extending tube life are also key elements, which the addition of EFB will produce. But folks who love their Fishers that much love it first for the sound it produces, and are very protective of it. Therefore, open driver stage surgery can easily make them nervous -- wondering if such a significant modification will change the sound of the unit. As a result, as stated above, the goal of this particular portion of the project is to mimic the performance of the original driver circuit (when operating as intended) -- warts and all -- so that the original sonic signature will be maintained. As to the warts of the stock design -- it is somewhat over compensated, which helps to produce that lush Fisher sound that Fisher is famous for. Therefore, the new driver circuit is designed to achieve that same end. The end result came out very well, as the pics provided will show.
More as it unfolds.
Dave
BELOW: These are 10 kHz square waves -- one from the new power amplifier design installed in Channel A, and one from the original stock design in Channel B. They are not perfectly identical, but these displays are more closely matched than many stereo amplifiers that (of course) have the same design in both channels. The new power amplifier design in Channel A is the top display:
BELOW: To show just how closely the two designs are matched however, this shot has one trace from above placed perfectly on top of the other, showing that the rise time, transient response, and flatness of frequency response are all but identical between the two channels, in spite of one channel operating with the old 7199 driver design, and the other with the new 7247 driver:
BELOW: Underside view showing the new driver stage design fully installed in Channel A (closest to the rear of the unit where the level controls are):
Last edited: