ef86 tubes, suitability in a mic pre, etc

[enthalpystudios]

i've been mulling things over in my head about a tube preamp using the ef86 and 6sn7 tubes that I have around.

because of (although untested by my own ears) reportedly poop distortions of the ef86 connected as a pentode, i'm told that it does ok as a triode.

what about as pentodes in pushpull? would it not then cancel out some of those distortions?

i'm trying to think about a different output stage since I don't have an output transformer suitable for the 6sn7's plate resistance.

I'm shooting for fairly high gain, 70-80ish total, probably 8-14 db of transformer gain, after factoring in both input and output transformers.

i just want one channel of it, and i have a 4 6sn7's and 2 ef86's to throw at it.

also, can some resistors be used to fake a centertap on a mic input transformer secondary to feed pushpull grids? or is this more of a trick suitable for heater taps?

thanks for humoring the tube noob questions.

billy

 

[NewYorkDave]

Hey Billy, I'm not trying to jump all over your sh*t, but...

because of (although untested by my own ears) reportedly poop distortions of the ef86 connected as a pentode, i'm told that it does ok as a triode.

C'mon, man, don't let Internet hearsay lead you around by the nose. You've got the tubes--try it yourself!

EF86 was designed for just that sort of application, the input stage of a low-level audio amplifier. It was used successfully as a pentode in a number of designs, both in Europe as the EF86 and in the States as 6267.

In terms of topology, here's one of many possibilities: EF86 as a pentode coupled to a 6SN7 (both sides paralleled) driving an output transformer, with overall feedback from the 6SN7 plate to the EF86 cathode. Idea #2: as above, but with second half of 6SN7 as a cathode follower, feedback taken from there.

 

[ciminosound]

The REDD47 used the EF86 in pentode mode. Seems to be popular. :green:

 

[NewYorkDave]

Yes, that's one of many.

Speaking of the '47, it wouldn't be a major job to adapt that circuit to use a 6SN7 for the output, in place of the E88CC.

 

[enthalpystudios]

not a bad plan, and thanks dave for not hesitating to jump on my shit ;]

very true about the hearsay and so forth. no try, no learn is always how it goes for me.

but yes, i do indeed like that plan.

i was reading a thread, where someone asked for a circuit you drew up quickly for rafa based on a 6sn7... it turned into a very likeable discussion involving prr concerning the ~10k Rp of the 6sn7, and he was talking about using an output transformer with quite a bit of primary inductance..... on the order of min 15k, but more preferable, it seemed.

anyhow..... maybe what i should try is to save up some bucks for the fairchild output clone that sowter makes, they're saying its 60kct:600, and I'd imagine that means plenty of turns and plenty of inductance on the primary, and probably a really sweet transformer as well.

with 1:10 input transformers, that'll basically mean no transformer gain though.

well to give my loadlines and operating points some practice, I'm going to give a go at drawing it up.

thanks

 

[NewYorkDave]

You don't need such a high primary impedance for a paralleled 6SN7 running, say, 15mA total plate current. Plate resistance of the two tubes in parallel would only be 3-4K. A 10K primary would be fine, and that's only 12dB of loss to a 600-ohm secondary.

 

[enthalpystudios]

heresay indeed.

10k it is ;] In fact I think I have an edcor of the 10k:600 variety. Now looking at the 47 schematic, it looks to me like the dc is all cap coupled at the output by C5, the 1uf there. but just to be sure, will the output transformer need to eat any dc? it looks like it won't, but no sense in not asking, as opposed to saving a poor transformers life.

 

[NewYorkDave]

The coupling cap blocks DC from the transformer primary. Use a cap that's rated for the full value of the B+, 300V minimum, 400V better. I'd use 2.2uF or even 4.7uF instead of 1uF.

The stock plate and cathode resistor values will work fine with the 6SN7. The closed loop gain will be a bit lower than with the ECC88; the open-loop gain of the 6SN7 in this circuit is only 20dB or so. Perhaps the best or at least easiest way to remedy this is to tweak the resistor values in the global feedback loop to compensate. (I specify "global" because there's also a local feedback loop around the output stage, comprising the 332K and 1.6Meg resistors).

Hint: the original used 1:7 input (200:9.8K) and 7:1 output (9.8K:200) transformers. You probably have no need to drive a 200-ohm line, so a 4:1 (10K:600) output is just fine--just bear in mind that your net voltage gain will be 6dB higher.

 

[Guest]

I used EF86 clones 6Ж32П both for mic and guitar pres as pentodes, they are excellent. In mic pres I used them with 4.7M grid leak resistor, no cathode bias resistors, 5V regulated filament voltage. If I remember right 220K in plate, 470K in 2'nt grid, but I may be wrong.


Hmmm... Now I am not sure: to delete this post, or to leave it as is... ePay reacts quickly...

 

[Marik]

[quote author="Wavebourn"]
Hmmm... Now I am not sure: to delete this post, or to leave it as is... ePay reacts quickly...[/quote]

Hahaha, not many here could even read 6Ж32П. :thumb:

 

[Kamel]

check the ampex mx 35 schema, maybe it is a good way to start
http://ftp.ampex.com/ampex/manuals/audio/mx35man/

 

[burdij]

Couldn't you drive one half of the 6SN7 as a cathode follower using the other half as a current source. This will allow you to use a low ratio coupling transformer. The pentode will provide more than enough gain for a microphone on its own.

Check the Radiotron book for some circuits with this configuration.

And contrary to other opinions, a pentode used in this application is rich in third harmonic and odd harmonic distortions versus a triode which, relatively speaking, is rich in even harmonic distortion. There are plenty of valid references to this phenomenon on the net.

 

[enthalpystudios]

or for high gain, perhaps I could build something similar to the 47 with the 6sn7 in place of the 2nd tube, and then use an extra 6sn7, which I've got, as a cathode follower and current source?

If that would let me use these a20's, which are 600ohm primary to 600/150/50, that would be awesome.

 

[burdij]

I can't right off hand find the Ro of a single 6SN7. In order to drive the 600 Ohm transformer directly, it may be necessary to use two sections in parallel as the cathode follower. Here is an article that discusses the tradeoffs and gives a method for calculating the appropriate load line.

http://www.tubecad.com/2005/June/blog0049.htm

 

[NewYorkDave]

A 6SN7 as a White cathode follower, biased for about 18mA of plate current, could drive 600 ohms fine. That'd be a much more efficient use of the tube than applying it as a conventional cathode follower.

RE: small-signal pentodes, in a mic preamp input stage, the voltage swing is generally so small that distortion is rarely a major problem. Also, most pentode-input mic preamps (for professional use) I've seen use negative feedback, I don't recall any at the moment that run open-loop. I'm sure you can dig up a couple if you look around; my point is that it'd be the exception rather than the rule.

Triode, pentode or whatever, noise is the biggest problem in input stages. Pentodes entail an extra noise source (partition noise) along with their higher gain. That's why pentode input stages usually employ special tubes selected for the purpose: EF86/6267, 5879, 1620 and so on.

(Speaking of 1620s, I'd love to get my hands on a couple of those and build myself a real 1940s-style preamp to use with ribbon mics. But I digress...).

At any rate, it seems reasonable to say that pentode-input mic pres don't necessarily suck, since they're all over a good percentage of our favorite old records! :wink:

Anyway, two 6SN7s in one mic preamp--that's a LOT of heater current for one channel of mic pre! Considering the tubes you have on hand, I'd be more inclined to use two EF86 with a single 6SN7 as a WCF, with a feedback loop around the whole deal. The REDD47 was interesting to invoke as an example circuit but there's no use in marrying yourself to the idea of building a near-clone of it this early in the game.

 

[enthalpystudios]

great thoughts.... in fact, I think I'm really into the idea of 1 ef86 and 2 6sn7s, despite the heavy power requirements. It seems just what I was thinking about.

I think a 6sn7 as white cathode follower into an A20 sounds like the perfect output stage for me, though... it's exactly the idea I originally wanted to try. It's a complex output arrangement for me, as far as correct implementation.

thanks again for the input you guys... I've started putting some of this together in superspice, not for the simulation, just because its free and has all the symbols... it seems, however, that I can't easily export a simple graphic or pdf.... is there a better free program just for simple schematic drawing around that I should be hip about?

I'd like to get something drawn up that can actually be analyzed...

thanks again

billy

 

[Guest]

I use LTSpice, but it has no damn transformers... :sad: