What about such a tube line driver? (oops! schem mistake modified)

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

pyjaman

Well-known member
Joined
May 31, 2007
Messages
295
Location
Poitiers,France
Hello,
linedriver.jpg


This is an idea for driving the output off a tube mic preamp (front end comming next).
The two halves off the tube drive the signal in the same voltage phase, but in opposite curent phase (same amount), current consumption remaining theoricaly constant.
This is an Idea I found in some John Broskie documentation (Tubecad).

What do you think I can expect from this design in terms of performances and drive ability ?
I know I could modify B+, Bias, and the two 10ks res. , but It'll be cool for My project to keep arround 20db gain for the stage...

thanks for any comment, analysis, and advices.

Laurent.
 
as drawn, you are taking your output from the B+, wich I assume is well filtered and quiet..?...so I assume that is a drawing error and not the design you thought up.
I also believe the feedback is positive, wich could be a problem. but the cathode to ground cap might take care of that?..
...other than that, I don't feel qualified/confident enought to answer...
j
 
> the two 10ks res.

There's only one 10K resistor shown?

The 10 ohm resistor shown in the second cathode pretty much ensures zero output.

> What do you think I can expect from this design in terms of performances and drive ability?

Can't guess.

Tube type is not specified.

Load is not specified (600r? 10K?).

The 220r and 0.1u do something at 7KHz..... what and why? Why does second cathode return to first cathode? It is indeed positive feedback, although with these values nothing can happen.

> current consumption remaining theoricaly constant.

Yes, it can. So? This lessens the need for a solid power supply, but V1 is eating large current without putting any of that to the load, "just" to make the power supply designer's work easier.

And if the iron is 1:1 and the load is 600r, then you do not have enough current for Pro Levels. Whereas if you paralleled both halves and got the full 30mA available to the load, or ran V1 at 2mA and V2 at 28mA, the signal ripple injected toward the power supply could be swamped with a 100uFd cap, which is not expensive today. And if the power supply is so poor that you feel you need to cancel signals, maybe it is so poor that you need wall-power ripple reduction too, which again leads to a big rail cap.
 
Oh, sh*t ! again I screwed up with my drawing software...
Sorry. fixed now.

Here is the description of this topology by the author :
constantdrawn.jpg


It explain much better than I could.

so the tube is a 5687.
Let's try some better values of resistors and a reduced B+:
linedriver.jpg


couldn't it have now better ability to drive  loads as low as 600R ?
anyway, V1's current consumption is still a waste. but any interest for this circuit?

Laurent.

 
I reckon the major flaw is using a 1:1 output xfmr. The rest may work. But I have some doubts about the linearity. If you want to make a grit box, why not, but it seems to be counteractive to the concept of reducing B+ induced distortion.
 
I've already built it (sans output transformer). Maybe I'll get around to measuring the response one of these days soon.

As for sound: not bad.
 
Hmmm....I think the designer got a bit too exited about this circuit. I can not see any real advantages. Perhaps it was just a think piece and never built or measured, not even simulated.  After a brief analysis in simulation:

1) Claim about the less problematic power supply design is somewhat true, BUT because the ridiculous current draw of the first stage it counteracts the advantage. Also, if you want to drive ANY LOAD, then the advantage is lost. Power supply in SE-amps should be good any way.
2) PSRR is not great. The explanation he gives is not true and in simulation PSRR is only 5dB. It gets _better_ by connecting the second stage to the ground instead of the first stage cathode resistor. The first stage is inherently sensitive to PS ripple, but the follower stage is not.  I cant see how they could cancel and they don't.
3) Linearity claims are dubious. The way the nonlinearity of a tube affects the distortion in a stage is not the same in grounded cathode and cathode follower (the latter being usually much more linear) I don't get this claim at all. Pulled out of the air. I was able to lower the distortion just by choosing the operating point of the first stage "right". Lower current, lower anode voltage.

Other things I'm not that exited about are the horrendous current draw. If you want your tubes red hot this is the way to go, but wasting that power in the gain stage is weird. If the topology would have the claimed properties then ok, but it doesn't.

The slight positive feedback is not essential in any way. At least not improving anything. Talk about "effectively bypassed cathode resistor" sounds great, but in this case doesn't do much. In this case you can go with the "normal" topology without a cap anyway and wont loose much gain.

Critical as always,

Jonte



 
You have 23mA per tube, 46mA total.

Yeah, it may drive 600r well, at gain near 10.

If you use the inherent leverage of the transformer, a 12AY7 and 10K:600 circuit could do similarly, with 1/5th the current drain. Which means 1/5th the power supply, or maybe 5 times the spare cash to improve the power supply.

I respect John deeply, but sometimes he wanders down paths which I don't believe are the most interesting.
 
Thank you, gentlemen, for these enlightening analysis, I was curious to understand a little more about this odd topology, but anyway, with such poor PSRR and waste of current, this is definitively not the way to go if the claimed goals are not even achieved.

Laurent.
 
http://www.tubecad.com/index_files/page0023.htm

look at the 2nd circuit - "Simple Hybrid Push-pull Amplifiers"

Looks like a "tube-transamp". I wonder if such a circuit could be made with a "Gainclone" chip (LM3875 or 3886). This could be sweet for amping  monitors etc. I wonder if such could work with substituting a FET for the tube.
 
> I wonder if such could work with substituting a FET for the tube.

Yes, it is possible to build electronics with no tubes at all.

Broskie has more recent essays about pushing power-chips with tubes; worth reading if only to show that it is not a simple snap-together.

> current consumption remaining theoricaly constant.

That's a problem in unbalanced work; but push-pull outputs naturally cancel supply current and crap. And since your load seems to be floating/balanced, why un-balance, use tricks to cancel current demands, then re-balance?

Why? Because 600 ohm loads need big current, and putting two low-current tubes in the path doubles the problem. Tubes and transformers go together! The transformer can couple a 5K tube into a 600 ohm load efficiently.

The top plan in conventional. If you can get a good 20K primary, this can make big line levels with under 25mA power consumption.

The bottom plan will also push large levels in 600 ohms, and cancel the supply, but is very brute force.
2zqvsl3.gif
 
prr,

in your opinion, would there be any point to a design similar to your sketch, but with a low ratio output transformer and less current?  125ma is not impossible, and in these days of cheap caps and power transformers, possibly cheaper than any decent output transformer.  but as you say "tubes and transformers go together" -- the issue is that a decent quality high impedance output transformer is an expensive item (relatively speaking), which makes the "conventional" plan unpopular. 

is there a compromise between these two examples that would make sense given the parts that are readily available off the shelf?  something that could drive a re-purposed power transformer, as in your 6EM7 headphone amp?  or is it folly chase such goals.

ed
 
edanderson said:
125ma is not impossible, and in these days of cheap caps and power transformers, possibly cheaper than any decent output transformer.  but as you say "tubes and transformers go together" -- the issue is that a decent quality high impedance output transformer is an expensive item (relatively speaking), which makes the "conventional" plan unpopular. 

I'm not PRR and I know you asked about using a re-purposed transformer with option two, but just thought I'd add that C.M.R.R. could be an issue if you used the pair of 6080's as is without a transformer.  Maybe not though and, in most situations, I find even an unbalanced out works just fine. 

Just throwing it out there  :)
 
PRR said:
> I wonder if such could work with substituting a FET for the tube.

Yes, it is possible to build electronics with no tubes at all.

Broskie has more recent essays about pushing power-chips with tubes; worth reading if only to show that it is not a simple snap-together.

Sure, true... but the one I linked has cathode feedback applied. Out of other Broskie's designs it looks closest to what OP posted ("closeness" is a very broad term, if we forget that the chip is inverting... :D ...and that in schematic there is a pic. of a triangle instead of a potato 8) )
 
Back
Top