Poseidon with KT-77s, or EC99s or 12BH7s

[EdStiles]

I'm rebuilding a set of Dyna Mk IIs, the original ones with the square circuit board. Purchased a set of Poseidon boards as well as a pair of SDS cap boards.

Question 1: Any thoughts about running KT-77s in this rig? I would assume that the bias would be set to 50ma / 500 mv per side? Any other mods?

Question 2: WHat about using a EC99 or 12BH7 in place of the 12AU7. On Ned's Triode Mk III board, both are listed as options (with associated circuit changes)

Question 3: See a post where someone used a 5751 in place of the 12AX7. any changes required to make that work (slightly less gain as I recall)

Question 4: Surge limiters, what size would be recommended for this amp with a 3.2A fuse?

Using Hexfeds (fast recovery) diodes in both bias and HV, my bias circuit has a 100 mfd and (2) 50 mfd caps, the latter bypassed with 0.1 and a 0.02 mylars. Using a 25k 2 watt pot to give me flexibility with El34s, 6550s or KT88/90s.

Thanks

[WA4SWJ]

Hi Ed,

I have some KT-77's but I have not fired them up yet. I do have three Mark III clones with poseidon drivers, an ST-70 clone with Shannon's driver board and an ST-35 clone with Shannon's board all with 5751 input tubes and 12BH7 phase inverters. No bias changes required. Just plug in the tubes. All of the amps sound very good. I recently made the tube change above to my ST-35 and the 5751/12BH7 combo significantly cleaned that amp up as compared to JJ 12AX7/12AU7 tubes (with the European ECC designations).

Hope this helps and good luck!

Regards,

[EWBrown]

KT-77s are pretty much interchangeable with EL34 / E34L / 6CA7, 50 mA bias seems about right. They sound great in an ST-70!

12BH7s shoud work well (or better) as 12AU7s, there may have to be a resistor change, I don't recall what was stated in Ned's driver board instructions. Not familiar with ECC99, can't help there. FWIW, 6GU7 is a 6V filament version of 12BH7, filament is pins 4 an 5, no "tap" on 9. 6CG7 / 6FQ7 not all that much different, more like a 9 pin version of 6SN7. Filament is also pins 4 and 5, no tap on 9.

IIRC, the "Ned Boards" have juper options for 12V or 6V filament tubes.

5751 should pretty much be a drop-in replacement for 12AX7, it has slightly lower gain factor , but no circuit changes should be necessary. Similarly, 5814A is a nice drop-in equivalent for 12AU7.

The European "ECC" numbers equivalents are ECC83, E83CC and ECC803 for 12AX7, and ECC82, E82CC or ECC802 for 12AU7. The 802/803 version are the "long plate" versions. There are "gold pin" versions, but they probably have more of a placebo effect than a real one...

HTH

/ed b in NH

[erichayes]

Hi All,

Having never worked directly with the Poseidon boards, Ed, I can't give you fersher answers, but the questions you ask are pretty generic--and I have been around the block a few times.

In my EL34 period, I liked to run them at (or they liked to be run at) 60 mA. The difference between 50 and 60 was analogous to looking at your back yard through the screen and sliding doors and pushing the screen door out of the way. Same should hold true for the KT77 only more so. A few years ago I moved up to (specifically, JJ) KT88s. Now I'm in the back yard.

Any place you have a 12AU7 and the option to use either a 12BH7 or a 6GU7 (heater current's the main issue, altho the 12AU7 is also shorter) go with the option. The 12AU7's OK until you've heard the 12BH7.

The 5751, again, is a drop-in sub for the 12AX7 as long as you have extra 75/150 mA heater current per tube available. The slightly lower gain can almost always be ignored as the circuit, rather than the tube, is dictating the overall gain. My favorite 5751 is the JAN Sylvania without the gold pins, altho the Sovtek is very impressive. Their problem is that they're fatter than NOS tubes, and are difficult, if not impossible, to stuff into a shielded socket. Don't waste money on the ElectroHarmonix 5751. It's the same tube with gilded pins.

I use a CL-190 in my 1773 amps, which draw around the same amount of current as a single MK II should, and haven't had a problem with any of them for the last 6 years. Mouser has them for around $1.25 in single lots (527-CL190).

Hope that's some help.

[Shannon Parks]

Question 1: Any thoughts about running KT-77s in this rig? I would assume that the bias would be set to 50ma / 500 mv per side? Any other mods?

Question 2: WHat about using a EC99 or 12BH7 in place of the 12AU7. On Ned's Triode Mk III board, both are listed as options (with associated circuit changes)

Question 3: See a post where someone used a 5751 in place of the 12AX7. any changes required to make that work (slightly less gain as I recall)

Hi Ed,

I second what everyone says here. The KT77s, 12BH7 and 5751 are all drop in replacements - no mods need for the board. The lower mu of the 5751 won't really affect anything, either. Tube roll away!

There are places on the Poseidon board to easily 'cut + jumper' to use 6CG7/6GU7 types for the LTP. Even V1 can be modded to accept a 6AX7. Anyone interested, just let me know.

Also, Eric mentions higher biasing - ie 60mA better than 50mA. I agree wholeheartedly. For all these PP amps running over 440V B+ (and within the tubes rating til 80%, let's say), I have always received much better test results on these power tubes. With a wealth of these tubes still available, I say don't starve a good tube. You'll still get at least a few thousand hours out of them. But keep within the tranny and tube (~80%) ratings. Just don't underbias.

Shannon

[EdStiles]

Thanks to all.

Got the 2 chassis back today (black powder coated), tomorrow I'll start assembly. I'll probably use some JAN 5751s that I have and the 12BH7s with KT77s biased at 60ma. Again, these are Mk II chassis so I have built a pair of extension plates to get the opening to match the Posideon boards, am mounting the boards on a quad of sorbane (sp?) vibration mounts. Inrush limiter will be a CL-190 and I should have plenty of power available with hexfed solid state retification for both HV and bias. SDS cap boards will replace the existing 500v 4 section cap with four sections of 50 mfd each.

Appreciate all the input, I'll let you know how it sounds. (Other parts of the system are CAL Alpha and Delta transport/DAC, either Marantz Model 7 or ARC SP3 with latest updates for preamp and KEF Reference 3 monitor with sub.)

Again, thanks to all

Ed

[Francois_G]

Let me tell you what I heard, or thought I, heard after a bit of tube rolling; and I'm not too concerned about using tube substitutes. So what I heard was a surprise to me.

I was running my MkIII (Poseidons) with nice/used side-getter 12au7 RCA in the second spot, with JJ12ax7 as voltage amp. It sound very good to me -- especially after replacing the Orange Drop coupling caps with some Russians (more about that in a later post after more listening).

Back to tube rolling -- I swapped the 12au7 RCAs with NOS RCA 5963's which are also a close 12au7 sub. Some folks feel they sounds better that the real thing. Well, they may in some applications, but not in my Poseidons. At least not immediately, and even after about 10 hours -- not too bad but a bit harsher and less enjoyable. I even tried them with a Telefunken 12ax7 without much improvement. When I put back the 12au7 RCA side-getters with the Teles things really sounded sweet.

Thoughts on what is going on here?

[erichayes]

Hi All,

Fran?ois, you have good ears and are willing to defend them. The 5963, as I mentioned a while ago in a different thread, is a highly specialized mutant of a 12AU7 designed to do two things: 1, Saturate with 0 volts applied to the grid and 2, Cut off when -15 volts is applied. Tube life was to meet or exceed the 10,000 hour criterion established for computer tubes. My first encounter with them was in 1961 where they were used as drivers in the Beckman/Berkeley decade counters that preceded Nixie tube readouts. Each of 5 5963s would fire a pair of neon lamps that would illuminate a digit etched on a translucent lens labelled vertically 0 through 9. Each decade was about 6" tall by about 1½" wide and 3" deep.

The last thing this tube was designed for was linearity, which is why it has such a transistor-like sound in audio applications. This is further exacerbated when the tube is run outside its idealized operating voltage of 150--what little linearity it has gets tossed in the dumpster.

[TomMcNally]

Eric ! You're tempting me into making an all tube clock !

http://mcnally.cc/clock.htm

Down with transistors and ICs ! Yellow_Light_Colorz_PDT_16

[erichayes]

Hi All,

Yeah, Tom, I had a feeling you'd get all hot and bothered over the Berkeleys. Unfortunately, I think you're going to have your job cut out for you; I eBay'd Beckman and Beckman/Berkeley for decade counters and got zero hits--for now.

Keep us posted if you're nuts enough to persue this convoluted holy grail.

[Shannon Parks]

My first encounter with them was in 1961 where they were used as drivers in the Beckman/Berkeley decade counters that preceded Nixie tube readouts. Each of 5 5963s would fire a pair of neon lamps that would illuminate a digit etched on a translucent lens labelled vertically 0 through 9. Each decade was about 6" tall by about 1½" wide and 3" deep.

Hey, this sounds familiar! Thermion picked up a HP decade counter at the Peoria Superfest last year (again in three weeks, gentlemen). Sold for $1 I think, but weighed 50 or 60 lbs. He pulled a ton of HP 5693s out of there and by golly, the Hickok pretty much fails them outright. I recall severe mismatch between halves.

[EWBrown]

I had tried a couple of NOS GE 5963s in the aforementioned Foreplay 2, and even though they didn't sound "bad" (they were better than the supplied Sylvania 12AU7s) they still didn't sound quite "right", so I rolled in a couple of presumably NOS RCA cleartops, which were very nice, and the latest "roll" is a couple of Telefunken 12AU7s pulled out of a junker FM tuner (I pre-tested them first and they measured as good as new).

The cleartops are going into the original FP3, for now. (not to be confused with it's not yet "born" 6SN7 cousin). I'd be hard-pressed to express any sonic differences between the cleartops and the 'funkens, I suspect that a "double blind" test with a better pair of ears than mine, would be necessary. Granted the 5963s were fresh out of their 40 year old boxes, and were not "broken in" very long, so maybe a few hundred hours' use might have sweetend them up, but I wasn't that patient...

/ed