Greetings to all!
I have in my possesion 2 DYNA A-451 output transformers (2200:4/8/16) in which Mr. David Hafler has written a particular amplifier design using a single 6AN8 driver and 4-KT88s in PPP config. per channel. My question is has anyone tried building an amp as Mr. Hafler had suggested and what are your impressions. This is a massive monolithic amp which requires a separate power supply and amp chassis. The circuit is very much similar to a MARK III where in fact the driver PCB can be utilized with just minor component changes. I would appreciate hearing from previous builders of this design and modifications they made would be a bonus. Thanks for your inputs.
regards,
edrel
I have in my possesion 2 DYNA A-451 output transformers (2200:4/8/16) in which Mr. David Hafler has written a particular amplifier design using a single 6AN8 driver and 4-KT88s in PPP config. per channel. My question is has anyone tried building an amp as Mr. Hafler had suggested and what are your impressions. This is a massive monolithic amp which requires a separate power supply and amp chassis. The circuit is very much similar to a MARK III where in fact the driver PCB can be utilized with just minor component changes. I would appreciate hearing from previous builders of this design and modifications they made would be a bonus. Thanks for your inputs.
regards,
edrel
> David Hafler has written a particular amplifier design using a single 6AN8 driver and 4-KT88s
He sold that commercially, with 8417s instead of '88s, as the Mark VI? I forget the number. I have one in storage. Mine sounds awful and I am sure it has a sick 8417 in it: they were not all very reliable tubes, and I do know this amp had a very hard life. The mono Dyna tube 120W seems to be a very rare beast, judging by online sales.
You already know how heavy the output iron is. The power transformer used by Dyna is about the same size, and those two lumps plus the hot quad of tubes is a BIG HEAVY amplifier. It is dangerous on the workbench: flip it over and you either break a switch or pinch blood from a finger. It was packaged as 19" rack-size, about 12" deep and about 9" high. It is bigger than a Bogen 300 watt eight-8417 PA amplifier... Dyna iron is BIG (and that Bogen was a real toaster).
The 8417 is a super-high-gain tube, so on paper you get more overall feedback than 6550 or KT88. In practice, you get grid-cathode shorts: I made a living with a dozen 8417 long-long ago and I would NOT use them in a new project. In fact if I ever restore mine, I will adapt the bias to run 6550, a workhorse tube that is reasonably available from several sources, new and old.
With 6550 or KT88, and Dyna iron, and Dyna-quality small parts, I really think the 120W will sound "just the same" as the well-known 35W and 60W amps, just 3dB louder. What little I could tell of my 8417 amp, between sick-tube bursts and bangs, sounded very Dyna, a hair less colored (at the 1 Watt level) than some of the others.
He sold that commercially, with 8417s instead of '88s, as the Mark VI? I forget the number. I have one in storage. Mine sounds awful and I am sure it has a sick 8417 in it: they were not all very reliable tubes, and I do know this amp had a very hard life. The mono Dyna tube 120W seems to be a very rare beast, judging by online sales.
You already know how heavy the output iron is. The power transformer used by Dyna is about the same size, and those two lumps plus the hot quad of tubes is a BIG HEAVY amplifier. It is dangerous on the workbench: flip it over and you either break a switch or pinch blood from a finger. It was packaged as 19" rack-size, about 12" deep and about 9" high. It is bigger than a Bogen 300 watt eight-8417 PA amplifier... Dyna iron is BIG (and that Bogen was a real toaster).
The 8417 is a super-high-gain tube, so on paper you get more overall feedback than 6550 or KT88. In practice, you get grid-cathode shorts: I made a living with a dozen 8417 long-long ago and I would NOT use them in a new project. In fact if I ever restore mine, I will adapt the bias to run 6550, a workhorse tube that is reasonably available from several sources, new and old.
With 6550 or KT88, and Dyna iron, and Dyna-quality small parts, I really think the 120W will sound "just the same" as the well-known 35W and 60W amps, just 3dB louder. What little I could tell of my 8417 amp, between sick-tube bursts and bangs, sounded very Dyna, a hair less colored (at the 1 Watt level) than some of the others.
I have not heard a Mark VI but if it uses the same driver board as a Mark III, I'd look closely at updating that when you decide to build an amp. There are much better driver circuits.
I like the idea of a Mark III driver board from Uncle Ned or Joe plus a quad of 6550s. I bet it would be rather nice.
rick
I like the idea of a Mark III driver board from Uncle Ned or Joe plus a quad of 6550s. I bet it would be rather nice.
rick
MARKVI/MARKIII
Thanks PRR,
MARK VI it was. The one using 8417 tubes but David Hafler designed a 120 watter different from Mark VI. As I have said this one's using 4 KT88's (or 6550s) and a single 6AN8 driver. Once I get home i'll direct you the DYNACO trans catalog for reference. If MARK VI would sound awful, I would think you are correct that the sound would be the same if not better than MARK III as they were using the same driver config. (and the same power tubes). Just hoping I would not pinch a finger in building one. ;o)
Hi RICKL,
Thanks for your input, and suggestions. By the way can you direct me to a download of that modifed MARKIII driver circuit you were talking about. I would think you can give me or you would know someone on who has built the modified MARKIII driver board assembly and their overall impressions on the improvement of the amp, at least for MARKIII. Thanks for your help.
regards,
edrel
Thanks PRR,
MARK VI it was. The one using 8417 tubes but David Hafler designed a 120 watter different from Mark VI. As I have said this one's using 4 KT88's (or 6550s) and a single 6AN8 driver. Once I get home i'll direct you the DYNACO trans catalog for reference. If MARK VI would sound awful, I would think you are correct that the sound would be the same if not better than MARK III as they were using the same driver config. (and the same power tubes). Just hoping I would not pinch a finger in building one. ;o)
Hi RICKL,
Thanks for your input, and suggestions. By the way can you direct me to a download of that modifed MARKIII driver circuit you were talking about. I would think you can give me or you would know someone on who has built the modified MARKIII driver board assembly and their overall impressions on the improvement of the amp, at least for MARKIII. Thanks for your help.
regards,
edrel
DYNACO TRANSFORMER CATALOG
the circuit of the 120 watter can be found here:
DYNACO TRANSFORMERS
the circuit is located on sheet 5 and the OPT table is at sheet 7.
regards,
edrel
the circuit of the 120 watter can be found here:
DYNACO TRANSFORMERS
the circuit is located on sheet 5 and the OPT table is at sheet 7.
regards,
edrel
MARK III MODIFIED DRIVER
Got the site and the downloads here:
http://www.curcioaudio.com/mk3upg_3.htm
regards,
edrel
Got the site and the downloads here:
http://www.curcioaudio.com/mk3upg_3.htm
regards,
edrel
> David Hafler designed a 120 watter different from Mark VI.
Different mostly because the Mark VI was produced years after that drawing was made.
Mark VI has a solid-state voltage-doubler rectifier with more C and less L. That makes power supply cost, bulk, waste, and weight much less obnoxious than doing it hollow-state with small caps and a big choke.
It has a time-delay relay, perhaps because he feared the start-up THUMP of a 120W amp would shred the speakers of that day.
The 8417 came out and appeared to be a better tube. It would be, except it pushed the limits of grid-winding. Some factories did it well, some didn't. Audio Research did a 8417 amp and it is said to work well with good tubes, throw sparks with less-good tubes. I had good luck with some 8417s, and also some spectacular melt-downs. A less excitable tube is probably a good idea. Any plan that makes 50-70 watts at 450V-500V with a pair of tubes can be adapted to a ~120W quad with these transformers. You could even run Pentode mode (ignore the screen taps) at 500V-600V and get 200 watts above 40 or 50Hz. My Bogens had smaller iron with less NFB and ground out musically tolerable bass down below 50Hz at the 180 Watt level.
The basic Dynaco driver is the same, near enough, in S-70, III, IV, and VI. Some differences in compensation caps, 6AN8 or 7199 depending on mood, few other details, and I don't know that they are physically the same board. But anything that applies to one can be adapted to all of them. I don't think it is a bad design: not sophisticated, but adequate. Parts quality could be better, and I know a hard-life Mark III's PCB will be toasted and brittle.
A radical idea is to triode-strap some 6550 or '88s. Get around 60W by my estimates. The Dyna driver will be real marginal: raise the supply voltage as much as you dare. A quartet or sextet of 300B would also be interesting, but driving them will be real hard work.
Different mostly because the Mark VI was produced years after that drawing was made.
Mark VI has a solid-state voltage-doubler rectifier with more C and less L. That makes power supply cost, bulk, waste, and weight much less obnoxious than doing it hollow-state with small caps and a big choke.
It has a time-delay relay, perhaps because he feared the start-up THUMP of a 120W amp would shred the speakers of that day.
The 8417 came out and appeared to be a better tube. It would be, except it pushed the limits of grid-winding. Some factories did it well, some didn't. Audio Research did a 8417 amp and it is said to work well with good tubes, throw sparks with less-good tubes. I had good luck with some 8417s, and also some spectacular melt-downs. A less excitable tube is probably a good idea. Any plan that makes 50-70 watts at 450V-500V with a pair of tubes can be adapted to a ~120W quad with these transformers. You could even run Pentode mode (ignore the screen taps) at 500V-600V and get 200 watts above 40 or 50Hz. My Bogens had smaller iron with less NFB and ground out musically tolerable bass down below 50Hz at the 180 Watt level.
The basic Dynaco driver is the same, near enough, in S-70, III, IV, and VI. Some differences in compensation caps, 6AN8 or 7199 depending on mood, few other details, and I don't know that they are physically the same board. But anything that applies to one can be adapted to all of them. I don't think it is a bad design: not sophisticated, but adequate. Parts quality could be better, and I know a hard-life Mark III's PCB will be toasted and brittle.
A radical idea is to triode-strap some 6550 or '88s. Get around 60W by my estimates. The Dyna driver will be real marginal: raise the supply voltage as much as you dare. A quartet or sextet of 300B would also be interesting, but driving them will be real hard work.
The Mk VI did get produced later - if it's the one with the "rack chassis".
The main problem with that amp is that it is unstable thanks to the rats nest wiring methods... (parasitics).
Actually, most of the Dynas are marginally stable, unless you keep all the compensations in place... which brings me to the next point: the Dyna "driver circuit" sucks. Period. Horrid.
Which, is odd, actually, since the predecessor, Acro, had some fairly nicely done input circuitry.
I'd skip the "Dyna" concept entirely with those nice big trannies, and go directly to a more sophisticated (read: better design) input stage and driver stage, possibly different output tubes too, and make urself an actually good amplifier regardless of where the power output ends up...
Oh, be sure to pick other tubes... 6AN8 is not high on my list, nor is the 7199 found in other Dyna offerings. I'd start by considering the 6SN7 and its relatives...
_-_-bear
The main problem with that amp is that it is unstable thanks to the rats nest wiring methods... (parasitics).
Actually, most of the Dynas are marginally stable, unless you keep all the compensations in place... which brings me to the next point: the Dyna "driver circuit" sucks. Period. Horrid.
Which, is odd, actually, since the predecessor, Acro, had some fairly nicely done input circuitry.
I'd skip the "Dyna" concept entirely with those nice big trannies, and go directly to a more sophisticated (read: better design) input stage and driver stage, possibly different output tubes too, and make urself an actually good amplifier regardless of where the power output ends up...
Oh, be sure to pick other tubes... 6AN8 is not high on my list, nor is the 7199 found in other Dyna offerings. I'd start by considering the 6SN7 and its relatives...
_-_-bear
Last Saturdays Experience
I spent all day Saturday "mucking about" with a prototype 120W monoblock using a Hammond 1650T O/P tranny - basically the same spec (2200 Ohm Raa) as the Dyna A451(but almost certainly not as good). I'm using 4 x Chinese KT88 Ultralinear connected.
Front end is a Cursio Audio Engineering PWB1A (Current Source Biased Cascode Diff Amp) direct coupled to 6SN7 Cathode Followers to drive the output tubes with 100K resistors from KT88 grids to bias supplies. That means the driver has to drive 50K load plus the O/P tube grid capacitances.
In this arrangement you need 110V pk-pk on the O/P Tube Grids for 128W out. Gain is a bit marginal to apply more than fairly small amounts of feedback so I put the amp back to open loop and tried 2 mods:
1) Pulled the JJ Tesla ECC88s from the cascode diff amp and put in Svetlanna 6N1P (mu = 43 vs 33) - this was dissaster because the cathode voltage of the lower valve in the cascode diff amp (ie collector of the current source) drops from about 3.8 volts to 1.9 volts and you therefore can't get enough input signal swing.
2) Put back the ECC88s and rewired the 6SN7 cathode follower drivers to common cathode stage 450V rail, 22K RL, 1K2 RK ( approx 8.2 mA Ia). This wasn't successful either because my -3dB point dropped from >70kHz to 19kHz. I tried adding balanced shunt feedback (220K) from the O/P tube anodes to the 1K2 cathode resistors of the 6SN7 driver - this dropped the -3dB point down to 17kHz ( I did'nt understand this) and so it was abandoned too.
My view from these experiments is that a cathode follower is essential to drive the output tube grids to any decent sort of frequency.
Also the CAE PWB1A circuit requires modification to return the bottom of the current source to a negative voltage rail in order to get any decent input voltage swing when using any tube except the ECC88.
Here is the schematic of my Down Under Designs DUD120
With the circuit as shown it produces 128W at clipping into 4 Ohms with Zout of 1.2 Ohms - CAUTION DZ1 on the schematic is drawn back to front.
It sounds good without being spectacular. Very strong and very fast. A bit too much distorting at very high levels. It might however give you some ideas for your project.
Cheers,
Ian
I spent all day Saturday "mucking about" with a prototype 120W monoblock using a Hammond 1650T O/P tranny - basically the same spec (2200 Ohm Raa) as the Dyna A451(but almost certainly not as good). I'm using 4 x Chinese KT88 Ultralinear connected.
Front end is a Cursio Audio Engineering PWB1A (Current Source Biased Cascode Diff Amp) direct coupled to 6SN7 Cathode Followers to drive the output tubes with 100K resistors from KT88 grids to bias supplies. That means the driver has to drive 50K load plus the O/P tube grid capacitances.
In this arrangement you need 110V pk-pk on the O/P Tube Grids for 128W out. Gain is a bit marginal to apply more than fairly small amounts of feedback so I put the amp back to open loop and tried 2 mods:
1) Pulled the JJ Tesla ECC88s from the cascode diff amp and put in Svetlanna 6N1P (mu = 43 vs 33) - this was dissaster because the cathode voltage of the lower valve in the cascode diff amp (ie collector of the current source) drops from about 3.8 volts to 1.9 volts and you therefore can't get enough input signal swing.
2) Put back the ECC88s and rewired the 6SN7 cathode follower drivers to common cathode stage 450V rail, 22K RL, 1K2 RK ( approx 8.2 mA Ia). This wasn't successful either because my -3dB point dropped from >70kHz to 19kHz. I tried adding balanced shunt feedback (220K) from the O/P tube anodes to the 1K2 cathode resistors of the 6SN7 driver - this dropped the -3dB point down to 17kHz ( I did'nt understand this) and so it was abandoned too.
My view from these experiments is that a cathode follower is essential to drive the output tube grids to any decent sort of frequency.
Also the CAE PWB1A circuit requires modification to return the bottom of the current source to a negative voltage rail in order to get any decent input voltage swing when using any tube except the ECC88.
Here is the schematic of my Down Under Designs DUD120
With the circuit as shown it produces 128W at clipping into 4 Ohms with Zout of 1.2 Ohms - CAUTION DZ1 on the schematic is drawn back to front.
It sounds good without being spectacular. Very strong and very fast. A bit too much distorting at very high levels. It might however give you some ideas for your project.
Cheers,
Ian
I would visit Curcio Audio and look around for a driver board. I'm using the Mark III upgrade board and it is a big improvement over the existing 6an8 driver board. Improved highs and better resolution.
I haven't used this but Triode Electronics also has a driver board for Mark IIIs. Uncle Ned's designs get a lot of praise but haven't heard an amp with his boards.
If I was going to spend time and money building a big amp, I'd look at gingertube's work and the driver boards from Uncle Ned or Joe. I really like 6sn7/6cg7 tubes. I've read good things about the 6N1P but haven't used one. You can decide if an octal or 9 pinner is your fav.
I'd look at KT88/6550 output tubes for availability and cost. You may want to start with gingertube's post.
rick
I think the 6cg7 that is used in the Mark III upgrade driver boards are 9 pin versions of the 6sn7 (see http://www.answers.com/topic/6sn7 .
rick
rick
> the 6cg7 ... are 9 pin versions of the 6sn7
12AU7 and 6CG7 are clearly the 9-pin descendants of 6SN7, but not exactly equivalent, The 6SN7 structure is a tight fit in a 9-pin bottle, so the 9-pins are re-designs.
The Williamson started with UK tubes, 6SN7 was close enough to work, but when you go to a 12AU7 a Williamson gets a little underbiased in the drivers. Some resistor tweaks are needed to bring it back to what Williamson intended.
The Mu of a 12AU7 drops at very low current. Lower than you would typically use in audio, but the effect is slightly useful in recording limiters. The 6SN7 shows much less drop of Mu at low current.
12AU7 and 6CG7 are clearly the 9-pin descendants of 6SN7, but not exactly equivalent, The 6SN7 structure is a tight fit in a 9-pin bottle, so the 9-pins are re-designs.
The Williamson started with UK tubes, 6SN7 was close enough to work, but when you go to a 12AU7 a Williamson gets a little underbiased in the drivers. Some resistor tweaks are needed to bring it back to what Williamson intended.
The Mu of a 12AU7 drops at very low current. Lower than you would typically use in audio, but the effect is slightly useful in recording limiters. The 6SN7 shows much less drop of Mu at low current.
Other CAE Offering
There is also a MKIV Driver Upgrade board at CAE.
SRPP (or it might have been cascode - been awhile since I looked)replaces the pentode section of the 6AN8 and parallel connected triodes replace the single triode section of the 6AN8.
It would certainly be an improvement over the 6AN8 circuit.
In my earlier post I stated that I abandoned common cathode 6SN7 as drivers for the KT88s because I lost too much high frequency response. Further tests have showed that the high frequency loss was NOT due to the 6SN7 CC stage not driving KT88 grid capacitance adequarely BUT RATHER my front end was not driving the miller capacitance of 6SN7s in common cathode adequately (gain about 11 and so 12 times the grid capacitance to worry about) - so disregard that information.
You'd think a guy with a name like mine would be able to sort miller capacitances easily!!!
Cheers,
Ian Miller - "Gingertube"
There is also a MKIV Driver Upgrade board at CAE.
SRPP (or it might have been cascode - been awhile since I looked)replaces the pentode section of the 6AN8 and parallel connected triodes replace the single triode section of the 6AN8.
It would certainly be an improvement over the 6AN8 circuit.
In my earlier post I stated that I abandoned common cathode 6SN7 as drivers for the KT88s because I lost too much high frequency response. Further tests have showed that the high frequency loss was NOT due to the 6SN7 CC stage not driving KT88 grid capacitance adequarely BUT RATHER my front end was not driving the miller capacitance of 6SN7s in common cathode adequately (gain about 11 and so 12 times the grid capacitance to worry about) - so disregard that information.
You'd think a guy with a name like mine would be able to sort miller capacitances easily!!!
Cheers,
Ian Miller - "Gingertube"
MARK III/IV UPGRADE
I have seen CAE upgrade for MARK III using 2- 6922s. This would be certainly one of my option. I am now on the drawing board laying out the parts and the chassis. I have 6550s and KT88s power tubes to try along with my 6SN7s as drivers. 6922s are not so expensive, i can get four of them soon..now my jensen imperials are waiting for this amp to drive them. certainly hope i can finish them soonest time possible!
I have seen CAE upgrade for MARK III using 2- 6922s. This would be certainly one of my option. I am now on the drawing board laying out the parts and the chassis. I have 6550s and KT88s power tubes to try along with my 6SN7s as drivers. 6922s are not so expensive, i can get four of them soon..now my jensen imperials are waiting for this amp to drive them. certainly hope i can finish them soonest time possible!
If you've got Jensen Imperials to drive, I'd recommend that you pick some triodes for the front/end driver tubes that are known to be "calm" like the 6SN7 or 12A_7 family and strap the outputs for triode... you don't need that much power and smoother sound will be a big benefit, imho.
BTW, lower DF (less feedback or none) may sound more better in the bass of the Imperials... something else to work with.
_-_-bear
BTW, lower DF (less feedback or none) may sound more better in the bass of the Imperials... something else to work with.
_-_-bear
thanks!- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.