Tubes circuits and lesser fidelity

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
eskimo said:
I think I understand most of this circuit, but what does R1, C1 and C7/C6 do?

R1+C1 are the zobel network for the input transformer. Transformers can ring in very high frequencies depending on loading conditions and transformer type. Sometimes audibly, sometimes only visible on oscilloscope and maybe negligible in real world use. Basically, square wave response (transients) might be skewed and you might want to tune that away. Most transformer spec sheets will tell you the optimal zobel, and sometimes only a resistor is needed. If you need to tune it yourself (difficult!) there are some very good threads around here.

And sometimes (most times?) zobel network isn't needed.

C7 and C6 are B+ smoothing caps, sometimes called "stabs", reducing ripple. These values are not critical at all, and anything between say, 10-200uF works. In fact they aren't needed for the amp to work. Depending on how "stiff" (regulated or capacitive) your B+ power is, they might not actually help at all.

[edit]

aww heck lassoharp answered these questions already. oh well the more the merrier.
 
Also put graphic EQs front and back, steep complementary curves.

+1.

Pre/post emphasis EQ is very, very important when designing nonlinear distortion circuits.
Generally speaking, boost high mids going in, and cut high mids going out. 20hz-20khz distorting circuits usually sounds like a 70´fuzz face pedal. (Tubes sounding marginaly better than bipolar parts)
 
Kingston said:
C7 and C6 are B+ smoothing caps, sometimes called "stabs", reducing ripple. These values are not critical at all, and anything between say, 10-200uF works. In fact they aren't needed for the amp to work. Depending on how "stiff" (regulated or capacitive) your B+ power is, they might not actually help at all.

Ah, of course! I'm so used seeing them in the psu section of the schem I got confused.
 
Here's what I sort of had in mind. The values are off since I'm pretty much clueless.
Two NYD stages. The two gain pots goes to a stereo pot for 1knob control. Output attentuator before the interstage trafo. Thoughts?
 

Attachments

  • dubbelmacka.jpg
    dubbelmacka.jpg
    116.1 KB · Views: 77
4 stages can be a plateful in terms of stability, distortion, phase shift etc. Without feedback its even tougher.

First thing I see is the A-18 most likely won't give you a low enough output impedance - you really want something like a 10 - 20Kish:600 here.

Second, I sense some overload problems at the grid of the 3rd tube.  Thing to remember is that 3rd = 1st tube in terms of it being designed to give max voltage gain to small signals(mic level).  By the time it gets out of the 2nd tube its a much larger signal (line level).  You may wind up having to keep the first stage volume control super low to prevent hard clipping at normal levels.

To do 4 stages without trouble would probably take a redesign. As is you'd be looking at some 90 or more dbs of gain.

For reference, here's one of only two 4 stage tube designs I know of: (note large amount of feeback)

http://www.groupdiy.com/index.php?topic=35357.0




 
If I wasn't clear, I wouldn't use the a-18 as an output but as an interstage between the NYD and the 436.
As for the 3rd stage, I left those values out, figuring the would rather be the numbers of stage 2 and 4, rather than stage 1. And yeah, 4 stages is a lot, i just figured it would make maximum use of the two double triodes.
 
Maybe you'd like something a tried and tested, instead of somewhat experimentally slapping four stages together like above?

Here's what I would suggest, these are all from NYD's plate of goods, but that's just because his designs are easy to understand and make sense from a builders point of view. Well actually it turns out the "two bottle" first stage is similar to some Gates mixer input stages.

1. Build the 6SN7 "one-bottle"
2. attach the below "second bottle" for additional 23dB of gain to drive your gentle 6SN7(EF86 works fine if you can't find 5879). Or just use the Altec "drive" you already have.
3. Use the pre-output-transformer gain attenuator to keep gain below transformer saturation. this is quite important because the above configuration puts out stupid amounts of gain.
4. experiment with lower plate voltages especially on the 6SN7 stages.
 

Attachments

  • NYD twobottle mod.jpg
    NYD twobottle mod.jpg
    17.1 KB · Views: 100
That doesn't look bad at all! Would sure be nice to try out.

When you say "altec drive" are you referring to the whole circuit or simply the first triode stage?
 
So, basically this?
I noticed the mod stage is identical to the 6SN7's first stage.
 

Attachments

  • NYD2bottle.jpg
    NYD2bottle.jpg
    104 KB · Views: 85
eskimo said:
When you say "altec drive" are you referring to the whole circuit or simply the first triode stage?

I thought you had already built the altec preamp. could as well drive the 6SN7 stages with that.

eskimo said:
So, basically this?
I noticed the mod stage is identical to the 6SN7's first stage.

Notice the first stage tube is a pentode wired as triode. The original draft by NYD shows the correct wiring, your new version doesn't.

Also remember to design the output attenuator.
 
Oh, right, forgot to connect 9-7 to plate.
As for the output att. would a simple pot work? I'm not too picky.
 
I don't know if a simple pot is a good idea. Might work, maybe others could tell how exactly.

I would instead do something like a 4-6 step(10dB steps) L-pad, matching it for the 10k impedance of the output transformer primary.

http://www.uneeda-audio.com/pads/

[edit]

here's a better calculator where you can just stick your impedance and decibel values. you can pretty much ignore watts in this circuit. 1/4w resistors will work.

http://www.bcae1.com/lpad.htm

PS. you could simplify the stepped pad by setting the series resistor to something like 8k, and use a 10k log pot for the parallel one.
 
grid leak bias on input.  lower bandwidth transformers.  lower operating voltage coupling capacitors.  open-loop pentodes.  really high value plate resistors.  there's plenty of stuff left which most people try not to do.  take a look at some of the older and cheaper tube tape machine mic inputs for some vibey ideas.  i think the ampex 601 runs an open-loop pentode and cascaded triode stages, for instance. 

you could try using blocking/high-value grid stopper resistors on grids of cascaded gain stages.  this way you can saturate the second one without changing the load the first tube's output sees as drastically.  This avoids a total meltdown on peaks but still sounds 'driven'.  The aiken amps site has a Norman Crowhurst (sp?) article posted about this though it's geared toward power amp output stages.

I agree, what you are after is not gross clipping.  The guitar amp analogy is right on.  More stages each contributing to the subtle distortion.  I think it's really hard to achieve this in one box.  I have better luck doing a little during tracking... bounce a few tracks around on tape... eq to make up for the crappy playback response of the record head... and so on.

None of these particularly apply to your specific circuit.  But maybe it'll give you some things to research/think about.
 
I'm finally getting around to building this thing. Gonna be a pretty hifi unit after all, but I hope driving it as hard as I want will get me a long way.

Doing the nyd 6sn7 circuit, copying the first stage and slapping it on up front. So three 6SN7 for a 2-channel pre. Probably gonna add grid stoppers to the original design. Any hints on the values on those?

Also, I wanna pad the signal before hitting the output iron, can I use a simple pot? I don't mind if it's a crude solution.
 

Attachments

  • PICT0001.JPG
    PICT0001.JPG
    178.2 KB · Views: 23
Eskimo,

You described it well, its like the difference between video and film only in sound.
But before you ditch your original Altec Pre with the cathode Follower read this article:-

http://www.freewebs.com/valvewizard1/dccf.htm

You can use that C/F to good advantage by arranging to operate near cut-off a' la Vox AC30/ Marshall amps, this does work.

You will need to use a chart for the 12AY7 (Franks Site) and adjust the operating point of the C/F to give you positive drive.

I agree with you that many of the changes do nothing, I found the same, but remember, the mic pre's used in Abbey Road decks (V72 and REDD47) all used plenty of feedback and they got the sound you want.  HiFi Geeks don't use feedback because they say its a purer sound, if that's true, then that's going in the wrong direction for you.  If you dump the feedback all you'll notice is the extra noise.  If you want colour then, use the C/F biased hot and it will give you some nice compression rather like tape does.

 
Thanks, that's an interesting read for sure!
The 12ay7 is long gone in my mind, and I've been set on a pair of nyd's for ages.

Interesting bit about getting some sort of compression and second harmonics out of it.

I've never thought of no feedback as a purer sound, but only as a way of getting and keeping lots of harmonic content.
Don't know what's the best approach really.
 
DaveP said:
but remember, the mic pre's used in Abbey Road decks (V72 and REDD47) all used plenty of feedback and they got the sound you want.

This is one of those odd myths that won't go away. V72 and especially REDD47 are awfully clean preamps and even the transformers they used in those were top of the line flat and THD-free. No distortion to be found there. Subtle "mojo" at best. Almost without exception the distortion and overdrive we hear on Beatles is tape. Slow tape with bad quality media they were still trying to perfect. Hello distortion and noise!

I can name some fifties rock records where I know the vocals have been tube distorted but even there I'm waiting to be proven wrong.

PS. Why did you have to post that direct coupled C/F link now that I've very nearly finished designing a guitar amp-like tube preamp/overdrive for tracking duties.  :'( I had no idea I could do that kind of asymmetric but controlled distortion.
 
Back
Top