The Mighty Bogen MO-200A
- Thread starter dcgillespie
- Start date
KT88 Fit
Double post.
Double post.
Last edited:
So, Dave- am I correct in deducing that the 6550 channel is winding up to be a lot like a non-UL Dynaco Mk II/Mk III circuit? If so, that makes me feel somewhat vindicated, being the guy who always seems to poo-poo replacing the stock driver boards on Mk IIIs... 
Yes, I understand you're putting EFB and such in there- but the fact that the driver sounds like a good ol' Laurent circuit, is somewhat satisfying for me. And, that also may greatly simplify my plans for modding out my Bogen MO60As, some time...
Regards,
Gordon.
Yes, I understand you're putting EFB and such in there- but the fact that the driver sounds like a good ol' Laurent circuit, is somewhat satisfying for me. And, that also may greatly simplify my plans for modding out my Bogen MO60As, some time...
Regards,
Gordon.
Hi Gordon -- If you've got MO-60s, then be sure to check out the earlier portion of this thread regarding my experience with the MX-60, and how the original driver circuit relates to it.
If you've poo-pooed replacing the stock Dynaco driver boards, I'm right there with you -- with the slight exception of the MK III board. The design is a good one to be sure. And you've done it a world of good by championing the adjustment of the screen resistor to optimize the inverter's performance.
I think this inverter is very good driving when 6BQ5, EL34, and even 6550 class tubes in the right scenario. In the MK III however, the circuit is running right on the very edge of capability. The inverter needs to output 110 volts peak-to-peak in the MK III, which requires about 420 vdc of B+ to the driver, which happens to be the value just available under full power conditions. But if the driver tube gets weak at all, or the pentode section isn't perfectly biasing the triode section, then performance suffers. It's a great circuit. I just think that in the MK III, it has to be watched carefully to make sure optimum performance is in fact being achieved.
In lesser applications like the MK II, ST-70, SCA-35, etc., there is enough extra drive available from the circuit that performance can still be good even with a weaker tube, or less than optimum inverter section biasing. I do favor the circuit. And it does work in the MK III. But unlike the other applications mentioned, it's not a plug the tube in and forget about scenario.
In the Bogen's application however, it is in many ways a non-UL MK III. One of the advantages of non UL operation with 6550s, is that all of a sudden, the Laurent circuit becomes very viable again (for me anyway). Peak drive requirements is reduced to 80 volts peak-to-peak, while the driver circuit installed can easily do 100 peak to peak volts as long as 400 volts B+ is maintained to the circuit. This allows for good performance in spite of the variables mentioned, and so becomes a very workable option in this case. One way the Bogen is not like the MK III however, is that each channel of the Bogen will very likely produce twice the power output of a MK III, while still using the same tube complement.
Dave
If you've poo-pooed replacing the stock Dynaco driver boards, I'm right there with you -- with the slight exception of the MK III board. The design is a good one to be sure. And you've done it a world of good by championing the adjustment of the screen resistor to optimize the inverter's performance.
I think this inverter is very good driving when 6BQ5, EL34, and even 6550 class tubes in the right scenario. In the MK III however, the circuit is running right on the very edge of capability. The inverter needs to output 110 volts peak-to-peak in the MK III, which requires about 420 vdc of B+ to the driver, which happens to be the value just available under full power conditions. But if the driver tube gets weak at all, or the pentode section isn't perfectly biasing the triode section, then performance suffers. It's a great circuit. I just think that in the MK III, it has to be watched carefully to make sure optimum performance is in fact being achieved.
In lesser applications like the MK II, ST-70, SCA-35, etc., there is enough extra drive available from the circuit that performance can still be good even with a weaker tube, or less than optimum inverter section biasing. I do favor the circuit. And it does work in the MK III. But unlike the other applications mentioned, it's not a plug the tube in and forget about scenario.
In the Bogen's application however, it is in many ways a non-UL MK III. One of the advantages of non UL operation with 6550s, is that all of a sudden, the Laurent circuit becomes very viable again (for me anyway). Peak drive requirements is reduced to 80 volts peak-to-peak, while the driver circuit installed can easily do 100 peak to peak volts as long as 400 volts B+ is maintained to the circuit. This allows for good performance in spite of the variables mentioned, and so becomes a very workable option in this case. One way the Bogen is not like the MK III however, is that each channel of the Bogen will very likely produce twice the power output of a MK III, while still using the same tube complement.
Dave
jaymanaa
RIP 1961-2018
I'm really enjoying this thread, thanks to all who have contributed. I was able to put together a decent stash of 8417s, but a matched octet is out of the question on my budget. This fact, and the fact that Micah's Bogen creation turned out so good has me looking at the Garter circuit for the outputs. I'll be watching here to learn more about the front end options.
rockur
Active Member
+1 :thmbsp:
Brick -- Thanks for following along! You blazed through this territory well before I did, and even though our approaches were very different, no doubt we both had to deal with many of the same common issues: Heck of a power supply, limited chassis options, and OPT challenges. Still, these amplifiers are incredible beasts which always makes for fascinating projects. Thanks for cutting a path, where there was none before --
Dave
Dave
6550 Version Now Complete
The new driver stage for the 6550 version has been installed now, with the NFB network adjusted for the same level of feedback that the original and modified version of the 8417 amplifier employs (17 db), as well as both the high and low frequency stability networks adjusted for similar results as well. Both the modified 8417 amplifier, and the new 6550 amplifier are utterly stable into any kind of load, with no amount of capacitance only load causing any tendency towards HF oscillation, while DC pulses produce very fast settling. Both of these qualities relate directly to reproduced sound, producing a cleanliness of HF sound without any shrillness, and tight bass that is well defined.
On the test bench, both amplifiers perform nearly identically, with the primary exception of input sensitivity. Since the original, common, screen grid supply and filtering circuit is still employed, each channel tends to regulate the other when each channel is driven individually. Under these conditions, each channel individually has a power bandwidth of 30 Hz, to 10 kHz at a sustained 125 watts RMS, and a power bandwidth of 20 Hz to 20 kHz at 70 watts RMS. At the 125 watt power level, distortion at 1 kHz is under 1%, with each channel producing about .8% THD at 100 watts RMS. The frequency response of the modified 8417 channel has previously been clocked at -1 DB @ 50kHz, while the 6550 channel can do a little better at -1 DB @ 60 kHz. Sensitivity wise, while the modified 8417 amplifier requires 1.75 vac to produce 100 Watts RMS output, the 6550 amplifier only requires half that amount at .87 vac. The target was to get this figure to around 1 volt, since that is more in keeping for equipment built for consumer oriented use, rather than the nearly 2 volt requirement this amplifier had for its original industrial intended use. Both channels are acceptably quiet on my 101 db efficient speakers, with most residual noise coming from the close proximity of power and output transformers on a chassis with a rather small foot print.
In the listening room, both channels sound identical for all intents and purposes. Both are incredibly responsive, with endless amounts of power on tap. It is really not possible to faze these amplifiers, as they laugh at the most complex recorded passages. On pipe organ music, the lowest bass notes are reproduced with great authority, with zero sense of work or strain. I ran the amplifier for 10 hours straight with the power transformer running very cool to the touch, and output tube bias remaining quite stable in both channels over the period.
Which channel is better? Tough question. Frankly, they're so close I would say that if you're so inclined, I'd build whichever is the least expensive for you to build. That will likely be determined based on your particular tube situation, or which tube you tend to favor most.
All that remains at this point then is to investigate the installation of EFB. It was hoped that possibly the 6550 version might not need EFB, with the inherently better screen efficiency those tubes have over the 8417, but that didn't appear to pan out in this case. As is, both amplifiers are very acceptable, with measured and sonic results running way, way ahead of the original amplifier as received.
A few pics to show where things currently are:
1. The 6550 channel is complete now, completing the original task of modifying the unit for stereo operation, incorporating a modified 8417 channel re-purposed for best home use performance, and providing a 6550 version to effect the same goals with more commonly available tubes.
2. A 10 kHz square wave from the 6550 version is virtually indistinguishable from that produced by the modified 8417 channel. Both have an acceptably fast rise time, quickly settling down to a very flat wave top, showing the very high degree of stability present.
3. Getting set up and warmed up in the listening room. This shot again shows the relative size of the two different sets of output tubes.
4. All wired up, ready to go. The clip lead attached to the screw at the bottom of the pic is a separate safety ground lead attached to the preamplifier. With this kind of power on tap, the last thing you want is a couple of hundred watts of hum blowing your ears and your speakers to kingdom come. This amplifier is definitely not in Kansas anymore.....
5. Another piece currently passing through the lab -- a Fisher 400 CX-2 stereo preamp is a nice mate for the Bogen. Together, the two made for many hours of some very enjoyable and effortless music.
So, with everything else complete now, EFB is the last stop of the last step of this project. It will require some reworking of the new bias supply installed, since at the time, a 6550 version of this amplifier was not being pursued. Now that it has been, it will require a slightly higher bias voltage for the EFB circuit to properly control 6550 tubes, in addition to the 8417s.
A bit of a break, but with a little luck, this project should be all wrapped up in another week or so. In the mean time, I should be able to post an updated version of the modified 8417 schematic, as well as publish the 6550 version schematic as well. The power supply schematic will be published after the EFB work has been finished.
Dave
The new driver stage for the 6550 version has been installed now, with the NFB network adjusted for the same level of feedback that the original and modified version of the 8417 amplifier employs (17 db), as well as both the high and low frequency stability networks adjusted for similar results as well. Both the modified 8417 amplifier, and the new 6550 amplifier are utterly stable into any kind of load, with no amount of capacitance only load causing any tendency towards HF oscillation, while DC pulses produce very fast settling. Both of these qualities relate directly to reproduced sound, producing a cleanliness of HF sound without any shrillness, and tight bass that is well defined.
On the test bench, both amplifiers perform nearly identically, with the primary exception of input sensitivity. Since the original, common, screen grid supply and filtering circuit is still employed, each channel tends to regulate the other when each channel is driven individually. Under these conditions, each channel individually has a power bandwidth of 30 Hz, to 10 kHz at a sustained 125 watts RMS, and a power bandwidth of 20 Hz to 20 kHz at 70 watts RMS. At the 125 watt power level, distortion at 1 kHz is under 1%, with each channel producing about .8% THD at 100 watts RMS. The frequency response of the modified 8417 channel has previously been clocked at -1 DB @ 50kHz, while the 6550 channel can do a little better at -1 DB @ 60 kHz. Sensitivity wise, while the modified 8417 amplifier requires 1.75 vac to produce 100 Watts RMS output, the 6550 amplifier only requires half that amount at .87 vac. The target was to get this figure to around 1 volt, since that is more in keeping for equipment built for consumer oriented use, rather than the nearly 2 volt requirement this amplifier had for its original industrial intended use. Both channels are acceptably quiet on my 101 db efficient speakers, with most residual noise coming from the close proximity of power and output transformers on a chassis with a rather small foot print.
In the listening room, both channels sound identical for all intents and purposes. Both are incredibly responsive, with endless amounts of power on tap. It is really not possible to faze these amplifiers, as they laugh at the most complex recorded passages. On pipe organ music, the lowest bass notes are reproduced with great authority, with zero sense of work or strain. I ran the amplifier for 10 hours straight with the power transformer running very cool to the touch, and output tube bias remaining quite stable in both channels over the period.
Which channel is better? Tough question. Frankly, they're so close I would say that if you're so inclined, I'd build whichever is the least expensive for you to build. That will likely be determined based on your particular tube situation, or which tube you tend to favor most.
All that remains at this point then is to investigate the installation of EFB. It was hoped that possibly the 6550 version might not need EFB, with the inherently better screen efficiency those tubes have over the 8417, but that didn't appear to pan out in this case. As is, both amplifiers are very acceptable, with measured and sonic results running way, way ahead of the original amplifier as received.
A few pics to show where things currently are:
1. The 6550 channel is complete now, completing the original task of modifying the unit for stereo operation, incorporating a modified 8417 channel re-purposed for best home use performance, and providing a 6550 version to effect the same goals with more commonly available tubes.
2. A 10 kHz square wave from the 6550 version is virtually indistinguishable from that produced by the modified 8417 channel. Both have an acceptably fast rise time, quickly settling down to a very flat wave top, showing the very high degree of stability present.
3. Getting set up and warmed up in the listening room. This shot again shows the relative size of the two different sets of output tubes.
4. All wired up, ready to go. The clip lead attached to the screw at the bottom of the pic is a separate safety ground lead attached to the preamplifier. With this kind of power on tap, the last thing you want is a couple of hundred watts of hum blowing your ears and your speakers to kingdom come. This amplifier is definitely not in Kansas anymore.....
5. Another piece currently passing through the lab -- a Fisher 400 CX-2 stereo preamp is a nice mate for the Bogen. Together, the two made for many hours of some very enjoyable and effortless music.
So, with everything else complete now, EFB is the last stop of the last step of this project. It will require some reworking of the new bias supply installed, since at the time, a 6550 version of this amplifier was not being pursued. Now that it has been, it will require a slightly higher bias voltage for the EFB circuit to properly control 6550 tubes, in addition to the 8417s.
A bit of a break, but with a little luck, this project should be all wrapped up in another week or so. In the mean time, I should be able to post an updated version of the modified 8417 schematic, as well as publish the 6550 version schematic as well. The power supply schematic will be published after the EFB work has been finished.
Dave
Attachments
Last edited:
20 Hz to 20KHz at 70 watts is still a boat load of power! Having tuned a few push-pull amps myself I know how hard it can be to get square waves looking that nice while still giving the fast rise times and the power bandwidth this one's got, especially with less than perfect output transformers. Pfweeew. (whistle). This thread makes me want to snag one of these amps, or one of similar pedigree, and do the same surgery you've done.
By the way, you mentioned the "Laurent circuit" a few posts back. Is that a reference to the frontend of the Dynaco Mk III, using the 6AN8?
By the way, you mentioned the "Laurent circuit" a few posts back. Is that a reference to the frontend of the Dynaco Mk III, using the 6AN8?
s-petersen
Scott
Hi Dave,
I have really enjoyed reading this thread also,Thanks for posting.
It may already be posted, but is there any good/bad points of using a quad of tubes per channel?and using lower taps for the speakers?as in the Eico st70 mod? If Bogen didn't need 8 output tubes,why did they waste the money?
Thanks,
Scott
I have really enjoyed reading this thread also,Thanks for posting.
It may already be posted, but is there any good/bad points of using a quad of tubes per channel?and using lower taps for the speakers?as in the Eico st70 mod? If Bogen didn't need 8 output tubes,why did they waste the money?
Thanks,
Scott
In the original design, it ran 4 per side very gently. With the way they had it set up, it could run 24/7 for years at a shot without eating the tubes up. It was basically just designed for a different purpose, where reliability trumps raw performance or engineering economy.
s-petersen
Scott
Thanks,Gadget.
Gadget -- Good eye, but in this case, that resistor is just worn looking ( not overheated) from previous service in various projects -- but it sure does give the appearance of being overheated in the pic. As for the caps, go figure. Like you, I've seen those things leak like a sieve, but these showed no leakage right up to their rated voltage, and have acted accordingly with complete composure throughout the entire project.
Kward -- Thanks for the kudos. Laurent is in fact the one credited with developing the original Dynaco amplifier front end (the MK II), and indeed the very concept of using a high performance pentode/triode tube with the pentode section direct coupled to the triode section, operating as a cathodyne inverter. With enough B+ supplied to the circuit, it allowed large tubes to be driven with the omission of the usual driver stage, with the pentode section making up for the lost gain, and the lack of an extra stage and coupling network providing a commensurate improvement in LF stability. Hafler recognized the benefits of the circuit, and teamed up with Laurent to form Dynaco. The rest is history.
S-Peterson -- Gadget nailed it. In re-purposing these amplifiers for home stereo use, it just made no sense to use four tubes where two would do -- and even then, the two used for home use will likely outlast the four originally used (per amplifier) as intended for continuous duty industrial service. While the final numbers haven't shaken out yet, the two tubes in each channel of the modified Bogen should end up operating at around 75% of their rated dissipation, which is conservative operation of the tubes -- particularly for the 6550s, based on a 35 watt plate rating. Also, with 125 watts RMS of power output on tap in each channel, it's scarey enough with that much power running around lose in the listening room, let alone 250 if you place an 8 ohm load on the 16 ohm tap. It could be done, but the complexity of design, and the heat given off by the thing versus everyday practical needs begs -- no -- screams the question, but why? It could be done, if only to show it could be done. And there is a rush knowing that kind of horsepower is on tap at a moment's notice. But honestly, for most tube enthusiasts, I think that's why 6BQ5 based amplifiers are so popular. Besides their sound, they also represent an excellent mix of efficiency, and appropriate power output, versus typical real world needs. Having 225 watts of quiescent power tube heat being dissipated to listen to what, 10, maybe 20 watts of power with inefficient speakers? Either that, or I'm just getting old, having already done such nonsensical things in the past. Yeah, that's probably it......
Dave
Kward -- Thanks for the kudos. Laurent is in fact the one credited with developing the original Dynaco amplifier front end (the MK II), and indeed the very concept of using a high performance pentode/triode tube with the pentode section direct coupled to the triode section, operating as a cathodyne inverter. With enough B+ supplied to the circuit, it allowed large tubes to be driven with the omission of the usual driver stage, with the pentode section making up for the lost gain, and the lack of an extra stage and coupling network providing a commensurate improvement in LF stability. Hafler recognized the benefits of the circuit, and teamed up with Laurent to form Dynaco. The rest is history.
S-Peterson -- Gadget nailed it. In re-purposing these amplifiers for home stereo use, it just made no sense to use four tubes where two would do -- and even then, the two used for home use will likely outlast the four originally used (per amplifier) as intended for continuous duty industrial service. While the final numbers haven't shaken out yet, the two tubes in each channel of the modified Bogen should end up operating at around 75% of their rated dissipation, which is conservative operation of the tubes -- particularly for the 6550s, based on a 35 watt plate rating. Also, with 125 watts RMS of power output on tap in each channel, it's scarey enough with that much power running around lose in the listening room, let alone 250 if you place an 8 ohm load on the 16 ohm tap. It could be done, but the complexity of design, and the heat given off by the thing versus everyday practical needs begs -- no -- screams the question, but why? It could be done, if only to show it could be done. And there is a rush knowing that kind of horsepower is on tap at a moment's notice. But honestly, for most tube enthusiasts, I think that's why 6BQ5 based amplifiers are so popular. Besides their sound, they also represent an excellent mix of efficiency, and appropriate power output, versus typical real world needs. Having 225 watts of quiescent power tube heat being dissipated to listen to what, 10, maybe 20 watts of power with inefficient speakers? Either that, or I'm just getting old, having already done such nonsensical things in the past. Yeah, that's probably it......
Dave
Schematics
As promised, here is the updated schematic for the 8417 version (only small changes made), and the new 6550 version with the 6AN8 driver section.
With the feedback level used and proper attention to LF stability, it allowed capacitive coupling to still be used between the pentode and triode sections of the 6AN8, with excellent LF stability still being achieved. This becomes more critical when a pentode gain stage is used, since the screen grid decoupling cap introduces an additional time constant into the circuit. Because of this reason, the vast majority of such designs use direct coupling between pentode and triode sections, to help negate the effects of the additional RC element at the screen grid.
Because capacitive coupling can still be used however, it means that the concerns for optimizing the pentode section's screen grid voltage is a moot point, since small changes in the operating point of the pentode section cannot affect the inverter section's bias. This then allows for very consistent performance to be had among the numerous 6AN8 examples I tried in the finished design.
For those wanting a somewhat lower input sensitivity (requiring greater input drive), an input level control can easily be added -- and in this case without any concerns for response issues that such controls often cause, since a pentode has no Miller effect. Therefore, varying input impedance levels have little effect on the HF response of the amplifier. Another option is to eliminate the cathode bypass cap, although this would likely require a touch up of the stability circuits.
In any event, there you have it. The EFB portion of this project is up to bat next, but that will develop rather quickly. With that, the final power supply circuits for both units will be finished, completing the circuit modification/development element of the project. Power supply schematics will then be published when that work is complete. Like the amplifier sections, they largely retain the original topology used.
1. The updated 8417 design. Only the DC balance pot value has changed to allow for easier adjustment of these very high Gm tubes. The input sensitivity figure has also been corrected.
2. The new 6550 design with 6AN8 driver. Besides the tube option this version provides, it is also more sensitive, is insensitive to variations in input impedance, and has a slightly higher frequency response. For all intents and purposes however, the performance between the two designs is virtually identical.
Dave
As promised, here is the updated schematic for the 8417 version (only small changes made), and the new 6550 version with the 6AN8 driver section.
With the feedback level used and proper attention to LF stability, it allowed capacitive coupling to still be used between the pentode and triode sections of the 6AN8, with excellent LF stability still being achieved. This becomes more critical when a pentode gain stage is used, since the screen grid decoupling cap introduces an additional time constant into the circuit. Because of this reason, the vast majority of such designs use direct coupling between pentode and triode sections, to help negate the effects of the additional RC element at the screen grid.
Because capacitive coupling can still be used however, it means that the concerns for optimizing the pentode section's screen grid voltage is a moot point, since small changes in the operating point of the pentode section cannot affect the inverter section's bias. This then allows for very consistent performance to be had among the numerous 6AN8 examples I tried in the finished design.
For those wanting a somewhat lower input sensitivity (requiring greater input drive), an input level control can easily be added -- and in this case without any concerns for response issues that such controls often cause, since a pentode has no Miller effect. Therefore, varying input impedance levels have little effect on the HF response of the amplifier. Another option is to eliminate the cathode bypass cap, although this would likely require a touch up of the stability circuits.
In any event, there you have it. The EFB portion of this project is up to bat next, but that will develop rather quickly. With that, the final power supply circuits for both units will be finished, completing the circuit modification/development element of the project. Power supply schematics will then be published when that work is complete. Like the amplifier sections, they largely retain the original topology used.
1. The updated 8417 design. Only the DC balance pot value has changed to allow for easier adjustment of these very high Gm tubes. The input sensitivity figure has also been corrected.
2. The new 6550 design with 6AN8 driver. Besides the tube option this version provides, it is also more sensitive, is insensitive to variations in input impedance, and has a slightly higher frequency response. For all intents and purposes however, the performance between the two designs is virtually identical.
Dave
Attachments
Hi Dave, is the input sensitivity you quoted in units of volts RMS or volts peak? I assume volts RMS.
Also, you are showing the core of the output transformer grounded. What is the purpose of that, and how is it practically done on the actual amp?
So there are three time constants on the 6AN8 version: pentode, first stage, and second stage. That's impressive you are still able to squeak 17 dB feedback out of this with three time constants and still have the amp be unconditionally stable! What stability margin does the amp have? Your write up implies a fair amount still, I take it.
Thank you. It has been a good thread.
Also, you are showing the core of the output transformer grounded. What is the purpose of that, and how is it practically done on the actual amp?
So there are three time constants on the 6AN8 version: pentode, first stage, and second stage. That's impressive you are still able to squeak 17 dB feedback out of this with three time constants and still have the amp be unconditionally stable! What stability margin does the amp have? Your write up implies a fair amount still, I take it.
Thank you. It has been a good thread.
Doesn't the core ground through it's mounting? I know power transformers sometimes have an electrostatic shield with a separate lead to shield the primary.
Thank you once again for your excellent work on this. The design stuff is rather beyond my skills and test gear compliment currently, but I'm learning thanks to excellent writeups such as this one.
Thank you once again for your excellent work on this. The design stuff is rather beyond my skills and test gear compliment currently, but I'm learning thanks to excellent writeups such as this one.