• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Long tail 6sl7 for Dynaco stereo 70?

I don't like that design. He's using equal plate resistors with a passive tail load. You can get away with that when using transistors, but 6SL7s have onehelluvalot less gain than any transistor. The output voltage won't be of equal magnitude unless you force balance by unbalancing the plate resistors. Even better is an active tail load to force constant current, and present the cathodes with a lot more effective impedance than the 47K tail resistor he does use.

As for drive capability, you need to pay attention to the Ci + Cmiller + Cstray of the driven stage if you are to aviod slewing at the top end. A 6SL7 may, or may not, be up to that task. If directly driving the grids of the finals, I'd say not too likely. If driving a cathode follower grid driver, or a subsequent small signal stage, then a 6SL7 could do that.
 
burnedfingers said:
http://www.aikenamps.com/LongTailPairDesign.htm

I was wondering if this would have enough guts to be used as the front end of a Dynaco Stereo 70.


Dude,

While the 6SL7 is a fine tube, it leaves (IMO) MUCH to be desired as a LTP. Slew limiting considerations indicate a high gm type as a LTP. The 'SL7 is high RP/low gm. :(

A 12AT7 as the LTP will yield as much gain as a 'SL7 would, but it has the requisite low RP/high gm.

The "classic" Mullard circuit with a 6GK5 as the voltage amplifier and a 12AT7 as the LTP is worth looking into. You get plenty of open loop gain and the high gm of both small signal types protects against slew limiting.
 
Miles Prower said:
I don't like that design. He's using equal plate resistors with a passive tail load. You can get away with that when using transistors, but 6SL7s have onehelluvalot less gain than any transistor. The output voltage won't be of equal magnitude unless you force balance by unbalancing the plate resistors. Even better is an active tail load to force constant current, and present the cathodes with a lot more effective impedance than the 47K tail resistor he does use.

As for drive capability, you need to pay attention to the Ci + Cmiller + Cstray of the driven stage if you are to aviod slewing at the top end. A 6SL7 may, or may not, be up to that task. If directly driving the grids of the finals, I'd say not too likely. If driving a cathode follower grid driver, or a subsequent small signal stage, then a 6SL7 could do that.

What's the Ci stand for in your post? The Cap between the 6SL7 LTP and the next stage?
 
I guess I can say this since I have all the 6CL6s I ever need. Using a 6CL6 LTP in pentode with a SS BJT CCS, I managed about 150 Vpp (single ended) out with <0.5%THD (my measurement limit) at 1KHz. Power supply was +340 and screen was 150V. Plate resisters were 12K and CCS was at 28mA. Single ended gain was 35V/V. It drove PPP EL34.


What it went in.http://www.audiokarma.org/forums/showthread.php?t=140121
 

tubewade

Member
2007-06-03 8:54 pm
Will it work? Yes. Will you be happy with it? Hmmm. I tend to agree with Miles that the anode resistors cannot be the same value if you want balanced output and that you really need a buffer between this stage and the output valves.

Also agree with Eli D. that the 12AT7 is better suited to the task. The Mullard circuit he recommends is proven and up to the task.
 

ray_moth

diyAudio Moderator Emeritus
2004-01-27 8:55 am
Jakarta
A 12AT7 as the LTP will yield as much gain as a 'SL7 would, but it has the requisite low RP/high gm.
Eli is right. However, you won't have very much gain with just a 12AT7 or, indeed, any double triode LTP as a front end. You need more stages to get enough gain for NFB. (Depends on how many holes you want to drill!)

For example, you could use a 6SL7 LTP splitter, followed by a 6SN7 differential driver (an all balanced design) which I like; or you could use a single-ended triode voltage gain stage, like a 6SL7 or triode-strapped EF86, followed by the 12AT7 LTP splitter à la Mullard.
 
Miles Prower said:
I don't like that design. He's using equal plate resistors with a passive tail load. You can get away with that when using transistors, but 6SL7s have onehelluvalot less gain than any transistor. The output voltage won't be of equal magnitude unless you force balance by unbalancing the plate resistors.

Randall Aiken said:
Note that in most cases, the out-of-phase output plate resistor will have to be made around 10% smaller than the in-phase output to achieve perfect balance between the two outputs.
:confused:
 
I use the same basic circuit in a 70 and it works Quite well.

BUT!!! Using much lower plate resistance tubes and different value plate resistors.
(Plus a driver tube in front of the splitter, no cathode bypass)

I use an E80CC for the driver tube to an 12BH7 for the splitter.
(many different tubes can be rolled in for the E80CC)

Yes you could do things to the tail, but I find this arrangement works well on a 70.

Something along these lines is the basic layout.
Cheers!

(Didn't look at the output phasing on there drawing, but wired like the drawing I posted it's fine.)
 

Attachments

  • ltp_12bh.gif
    ltp_12bh.gif
    5.9 KB · Views: 413
I agree with these guys.  The 6SL7 is an outstanding performing and sounding device when used within its demesne, but maybe this is not it.  Plate resistance is way too high, and you're going to divide all your signal away in the grid circuit, or else the top end is going to roll off due to Cinput like Myles said.  Your basic circuit is something I've been developing lo these fifteen years, and for the EL34, this is my current recipe:

[IMGDEAD]https://www.audiotropic.net/Demo/smallamp34.gif[/IMGDEAD]

The 6GK5 is a 'VHF amplifier', it was used in teevees somewhere.  It has all the chars to do the job (plate resistance of a 6SN7, in-circuit gain of a 6SL7, great linearity, juicy tonality), single triode so it sidesteps the sections-in-a-bottle matching headache, and, being a forgotten teevee tube, the vendors will beg you to take them at a dollar each.

This is not mouth music on two accounts; I have the breadboard built and it is maybe the best sounding amp in my checkered career, and I have over three hundred 6GK5s tested and matched on my shelf for my future delectation. And my clients'.

One geek's opinion, YMMV, et dissing cetera.

Aloha,

Poinz
 
sorenj07 said:

Take a look at Aiken's Original Schemo

Next, break it apart into Two Separate Circuits

Next, ask what happens. The Zo of V1A is equal to: r(k) || Rk. Given a 6SL7 and the nominal characteristics: r(p)= 44K; g(m)= 1.6mA/V. That gives an r(k)= 1K07 since it's in parallel with the tail resistor. The input side of V1B is going to have an r(k)= 1K77, since this side has that Rp to consider. That's going to give a voltage division of ~0.62, not the expected 0.5. Of course, the imbalance just gets worse the lower gain VTs you use for an LTP. Or why making the tail load as large as possible with a CCS improves the operation greatly.

Sure, you can compensate somewhat by using unbalanced Rp's, but that can have consequences at the higher frequencies since the Zo of each phase is no longer equal. This is why I opt for active tail loads for LTPs.

Compare the r(k) of this type with the r(e) of a BJT operating at a very modest Ic= 1.0mA, which would be: 26R. This would correspond to a g(m) of 38.5mA/V -- a value that very few VTs, even high gain pentodes, can come close to. Given that, it's easy to see why a BJT LTP doesn't have much problems with imbalance even for tails that appear "short". After all, a 2K6 tail resistor would be two orders of magnitude greater. Of course, a BJT LTP will benefit from a CCS as well, and one should be included, especially if emitter degeneration is also applied.
 
This is where the ~10% differential between plate resistors mentioned by Aiken comes into play.

I bought a driver board that had (3) 12AT7's for my last Dynaco Stereo 70. I believe it had a 12AT7 with long tail and plate resistors that weren't the same value plus a pot to fine tune it.

I had thoughts about trying to make something myself using a 6SL7 because I like the sound of the tube.

There is also a circuit by Mapletree that uses a 6SL7 in long tail I believe. I could be mistaken however.

Anyway I appreciate the comments and the chance to learn.
 
burnedfingers said:
I had thoughts about trying to make something myself using a 6SL7 because I like the sound of the tube.

No reason not to use them, as the 6SL7 has some outstanding linearity, and does sound very good. Much better than a lot of the mini's that replaced it. It's just a question of using them right.

There is also a circuit by Mapletree that uses a 6SL7 in long tail I believe. I could be mistaken however.

Anyway I appreciate the comments and the chance to learn.

Here's how I used a 6SL7 LTP: Vixen Main Schemo. Here, I included active tail loading with cascoded BJTs to make a really good CCS. The design also includes grid drivers that don't make a hard load for the 6SL7s. 6SN7s have small internal capacitance, and you don't have Cmiller to deal with and the cathode follower "reverse bootstraps" the Cgk to a much smaller effective value. That way, you don't need a lot of plate current, and an Ip= 0.575mA complies with the "Rule of Five" at 30KHz, so there's no problem with slewing. Of course, connecting the 6SL7s directly to the grids of the 807s would make for onehelluvabig slew problem, not to mention what would happen during a transient overdrive that takes the 807 grids positive.