need help with cathode follower circuit

[Ben N]

B+ is your plate power supply, which is what goes ionto your amplified signal, and it positive DC (hence the + sign). It is completely separate from the heater supply, which can be AC or DC. Back when the most common tube circuits were battery powered radios, they had three batteries for heater filaments, plates and bias, and they were designated A, B & C. That B+ designation is a holdover from that time, even though most tube circuits today are posered by a transformer/rectifier/filter arrangement, and generally draw all of their power from separate secondaries on a single power transformer.

No offense, but your questions suggest that you should do a little basic reading about tubes and tube curcuits before you take on a project.

[REGNAD]

thanks, that all makes sense. Ive done the research but all i have found uses terminology that wasn't clear to me, so i posted these questions. i think im good now! with this circuit anyway. im going to the garage to build it now and see what happens...

[REGNAD]

assuming i understand the "given" connections i built this circuit. didnt work though, a little noisy on its own, and it augments the noise of other fx. im short on power supplies so i tested with a 12vdc on heater and 9vdc 600mA on B+... could this be my noise problem?

[Ben N]

This is where I get (way) out of my depth, but it seems to me that 9v is very low for any 12a_7 plate. Your schematic shows 24v; most "normal" 12a_7 circuits get 100+ volts. Aside from whatever other challenges such a low "starved" plate voltage might raise in terms of biasing and headroom, I would imagine that it also significantly degrades your S/N ratio. I doubt very much that Cornish is powering his buffers off anything less than a full-blown amp-style power supply, with something like 180-220v on the plates.

There are some recent/current threads, most of them initiated by forum member frequencycentral, that address providing adequate plate voltage in tube pedal circuits using voltage doublers, charge pumps, etc. They might be of use to you.

BTW, the old warning can never be repeated too many times: before you start even think about working with high voltage, make sure you know WTH you are doing, because ignorance could cost you your life.

[REGNAD]

ok, ill try a full 25 volts and go from there...

[punkin]

I wonder that if working with a low voltage like that how a 12U7 might sound in this application? Of course the components would change a bit ... Just a thought.

[Gus]

Look at the curves of the 12au7.
http://www.mif.pg.gda.pl/homepages/frank/sheets/093/1/12AU7A.pdf

What are your voltages at the tube?
Cathode to ground VDC?
33K and the 330 ohm node to ground VDC?

How weak is your output drive?(current in the cathode leg and ohms law)

Do you want that high an input resistance?

Using DC on the heater can be tricky you need to understand about how power supply noise can couple from the heater.  I would use the 6.3VAC or 12.6VAC wiring for the filament AC can be quiet if done correctly and DC can have added noise if not done correctly.

[REGNAD]

i have no idea what im looking at with the curves :\
How weak is your output drive?(current in the cathode leg and ohms law)<<<<is there another way to word the question?
Do you want that high an input resistance?<<<<i have no idea
will a simple ac plug in wall wart work for the AC heater power supply?

[Ben N]

REGNAD, meet Gus. Gus (Augustus Smalley) is one of the Wise Men of DIY, worthy of great respect and from whom one can learn a great deal. Rather than posting pat answers and ready-made solutions, however, he provides clues and guideposts to encourage folks to learn stuff from the inside out. Sometimes the duller among us (like me) are left scratching our heads.

[DougH]

If you really want to understand how followers work, I suggest looking it up in "The Art of Electronics". They have a good writeup on emitter followers (and maybe source followers too, don't remember). They have nice explanations of input/output impedance and biasing with followers. Emitter/source/cathode followers all operate on similar principles btw.

edit: (Oh never mind... I see this kind of stuff was already covered earlier in the thread.)

[gmoon]

Well, let us know how well a 12AU7 works as a follower w/a low plate voltage. Other starved-cathode designs are courting non-linear response (which theoretically isn't gonna happen with a CF.) Maybe it will work OK, maybe not.

Power-supply noise is always an issue with the low-voltage tube projects. Read the Valvecaster thread for advice.

The other thing that pops out is 24V for the filament...but that's just a typo, right? 

I hate to "pile on," but if it works correctly, it's not going to sound different than a well-designed solid-state follower. Oh, well. Not a reason by itself to not build. Of course, a haphazardly-designed follower will effect the frequency response, and you might actually like that (kinda defeats the purpose, though...) Good luck.

[DougH]

Maybe some of the "mojo" stuff comes from the claims that CF's in Marshall preamps compress a little. I have not tested this myself, and in fact used MOSFET followers in a couple amps I have built and am happy with them. But some do claim that a tube CF in a preamp like this will compress somewhat and change the sound a little. Of course, this is with a much hotter signal, after two or three gain stages. I doubt you're going to have this behavior with a passive pickup level signal which would be about 1V p-p at most, usually.

But the problem, as I've stated many times before, is that people always want to know how to build "widget X" because they have a built-in assumption that widget X is going to do something to their sound that they want. What they need to ask is "how do I do that something to my sound?" and then let people suggest the appropriate widget to use. When you ask how to build something without talking about what you specifically want it to do for you, people in a forum like this will make all kinds of assumptions about what you expect it will do for you and then you get a lot of confusion.

[morcey2]

I go away for a week or two and things get exciting.

Maybe some of the "mojo" stuff comes from the claims that CF's in Marshall preamps compress a little.

The 'mojo' from the CF in  marshalls comes as a side-effect of a direct-coupled CF, not an AC coupled one like this.

What are your voltages at the tube?
Cathode to ground VDC?
33K and the 330 ohm node to ground VDC?

With that setup, the cathode voltage should be right at about 12-13V.  I've simmed it, but I haven't built it.  I have built other stuff simmed with the same software and models which were reasonably close to real life measurements.

How weak is your output drive?(current in the cathode leg and ohms law)

The output impedance for this setup approximately 1/gm (gm being transconductance).  The gm of a 12AU7 at 24V plate voltage is (guessing because of the low voltage) 1600-µMhos.  1/0.0016 = 625Ω output impedance.  That should be able to drive just about any load.

I wonder that if working with a low voltage like that how a 12U7 might sound in this application? Of course the components would change a bit ... Just a thought.

My guess is that the characteristics of a 12U7 are so close to a 12AU7 at that voltage that nothing would have to be changed.  

I personally still think it's more trouble than it's worth to do this with tubes just to say that it's tube driven.  Again, it's very doable, but won't gain you much over other types of buffers.  

Matt

Edit:  Also, on the heater part of the schematic that was added, the voltages should be either +12 and 0 or +6 and -6.  If you run them as shown (+12 and -12) you'll burn out the heater filament in very short order at 24V. 

[Gus]

morcey2

I don't trust sims or even the older drawn curves of tubes at low voltages and current.

Lets say the sim is correct and the tube cathode sits at 12VDC

330 + 33k = 33.33K        12VDC/33.33k = .36ma

.36ma x 330 ohms = about .12VDC bias

If you look at the curves http://www.mif.pg.gda.pl/homepages/frank/sheets/093/1/12AU7A.pdf it is hard to tell what is going on at the voltages and currents.

I find when operating tubes at low currents it is better to measure.

Lets take the simple EF, 9VDC, 10K emitter resistor, emitter at 4.5VDC (or higher because of what follows it)   4.5VDC/10K =.45ma for higher input R bootstrap it.

Look at the EF section of this it might make an nice buffer you can even make it work at higher voltages.  The need for the bootstrapping becomes less if you increase the emitter resistor value as the supply voltage goes up
http://www.aronnelson.com/gallery/main.php/v/gus/wha1.gif.html?g2_imageViewsIndex=1

[DougH]

The 'mojo' from the CF in  marshalls comes as a side-effect of a direct-coupled CF, not an AC coupled one like this.

That's true. Which means there's even less reason to build a tube vs. ss buffer for this.

[morcey2]

The 'mojo' from the CF in  marshalls comes as a side-effect of a direct-coupled CF, not an AC coupled one like this.

That's true. Which means there's even less reason to build a tube vs. ss buffer for this.

HEATHEN!!!!   Oh, wait.  Wrong board.

[REGNAD]

what im after is to be able to isolate each of my fx pedals so that the behave as if they are the only effect in the chain. i have a compressor, bk tube driver, and big muff among other less noisy fx. all are nice and quiet by themselves, but when two or all three of these boxes are on my noise floor gets all kinds of angry! i like the idea of using a tube and understand that a buffer circuit would give me this isolation... a full schematic showing all connections would serve me better, i dont know the "understood" connections. im using a 12vdc power supply for the heater. in the systems i service and interact with lable things that way. its simply the posative and negative connections off the wallwart, not 24 volts.

[Ben N]

OK, as long as those pedals are in series, buffering wont change the accumulation of noise that you get when turning them all on. If anything, adding multiple buffer stages will slightly increase the noise. If you really want to lower your noise (without using a noise gate), you need to go into each pedal and tweak for minimal noise, e.g. subbing metal film resistors for other types, lowest noise active components, improving shielding, etc. But know that each of the pedals you mentioned is inherently noisy, and you will only be able to reduce it so far, plus there may be sonic costs, such as going with low-noise, high gain transistors in your Muff. This should improve your S/N ration, but, depending on what you are after (e.g. Ram's Head, etc.), may not do much for your tone.

[REGNAD]

why is it that these 3 pedals "feed" off each other? by themselves they are almost completely quiet. also, how do you tell which components are the cheaper variety?

[morcey2]

why is it that these 3 pedals "feed" off each other? by themselves they are almost completely quiet. also, how do you tell which components are the cheaper variety?

If I hookup my zombie chorus and PT-80 at the same time, I get a really weird high-frequency buzz.  I think it's got something to do with the HFO clock thingies in both of them.  (That's as technical as I get on a monday night.)

Where is the compressor in the chain?  If it's last, it'll boost the noise from the other two when there's signal to speak of coming from the guitar.

Matt