ECC832 Driver and Conertina

[lowell]

hey all,
I'm driving a 2 6L6 and 26V6 p/p output section from a ECC832 (12dw7=12au7 and 12ax7) tube. FYI the grid leaks for the 6v6 are 470k cathode bias, and grid leaks for 6L6 are 100k fixed bias. Also all coupling caps in between are .047uf.

The concertina in the 5E3 uses 56k Ra and 56k/1.5k Rk but that amp also used a standard 12ax7 tube. I'm curious what you all may recommend for values when the concertina is a 12au7. The Ra of a 12au7 is listed as 7k! Far far lower than the 12ax7's 62.5k. My circuit is showing these voltages w/ the 56k's.

p 212
g 70
k 120

I'm thinking I could go lower than 56k to get more current and set the bias right in the middle.

I realize that lower values will allow for more current capability and potential smoothing out of grid limiting on the power tubes' grids, but I'm green when it comes to clipping characteristics of the concertina itself, especially the cathode output. I have a generator and scope and will definitely be doing some experimenting but thought I pick yalls brainzzzzzz.

[lowell]

damn i mispelled concertina in the title... can't figure out how to edit that.

[Chuck H]

Yes, the much lower plate R is standard fare with 12au7 tubes. Look at some old Ampeg schems and you'll see. But I've never seen the plate R's as low as 7k or even teens.

Now you have me curious. You and I both need to look up some load lines for 12au7's and see about the optimum bias for that tube. It really is a different animal than the 12ax7 even though alot of people just seem to know it as a lower gain tube for a 12ax7 socket.

Chuck

[Merlinb]

The current is largely irrelevant, what you're interested in is voltage swing. 6L6s are hard to drive, so if you want plenty of PA overdrive you need lots of swing. If you don't, then you don't need lots of swing and you can use smaller resistors on the concertina.

56k+56k will give you a LOT of swing from an AU7, like 120Vp-p to each 6L6, assuming your HT is in the 300V region. Maybe that's what you want, in which case 47k to 56k is a suitable choice. If you want a lot more headroom then values as low as 10k+10k aren't out of the question...

[lowell]

Yes, the much lower plate R is standard fare with 12au7 tubes. Look at some old Ampeg schems and you'll see. But I've never seen the plate R's as low as 7k or even teens.

Now you have me curious. You and I both need to look up some load lines for 12au7's and see about the optimum bias for that tube. It really is a different animal than the 12ax7 even though alot of people just seem to know it as a lower gain tube for a 12ax7 socket.

Chuck

As far as I can see it looks as though where I have it I will get grid current limiting around 50v p-p. Now, when the cathode clips is the mystery to me. I'm not sure how to do a load line for that... or am I. I'll report back after some brain storming. Calculating gain I may be able to see when the anode will clip on positive and negative cycles.


Also, merlinb you speak of voltage, but doesn't the concertina provide less than 100% of the input voltage? I'm not clear on this, but if that IS the case then why not bias it for maximum current? Unless that is you want earlier power tube grid conduction?

[Merlinb]

As far as I can see it looks as though where I have it I will get grid current limiting around 50v p-p. Now, when the cathode clips is the mystery to me. I'm not sure how to do a load line for that... or am I. I'll report back after some brain storming. Calculating gain I may be able to see when the anode will clip on positive and negative cycles.

Just draw a normal load line equal to Ra + Rk, then bias accordingly- there's no trick to it. If you want max headroom in the concertina, use centre biasing.

Also, merlinb you speak of voltage, but doesn't the concertina provide less than 100% of the input voltage?

Yes it has gain less than unity, so if you want 120V out, you need to put 120V in (well, a little more, but that's not a big deal).

I'm not clear on this, but if that IS the case then why not bias it for maximum current?

Well my answer to that is, what is that current supposed to achieve? Why do you want it? Having a lot of current flowing in the concertina doesn't really do anything for the PA. Ok, it might improve the slew rate, but for a guitar that really isn't an issue...

[lowell]

As far as I understand it, the lower the output impedance of the PI the more rounded the clipping will be if/when the grids of the power tubes conduct.

[Merlinb]

As far as I understand it, the lower the output impedance of the PI the more rounded the clipping will be if/when the grids of the power tubes conduct.

Unfortunately, output impedance is only half the story. (For examplem an ECC83 cathode follower has an output impedance of a few hundred ohms. But does that mean it can drive even 10mA into a load? Probably not.)

But second, you're power valves will (should) have grid stoppers, in which case the current you can get into those power valves is so puny that quiescent current in the driving stage is not especially important. For example, if you use 5k grid stoppers, you need 40V just to drive 10mA into the grid! To drive a few tens of milliamps, which is what you'd need to make any significant difference to the 'softness' of clipping you'd need hundreds of volts of swing from the PI, in addition to the swing needed just to get it to the point of clipping in the first place!
What I'm trying to say is, quiescent current in stages is of little important in guitar amps. Hifi amps maybe, but guitar amps, don't worry about it.
By all means play with load values- its not trouble to tack in different values, without hurting the valve.

[lowell]

"The rate at which the signal is clipped also depends on the source resistance, since
the input voltage is dropped across this. If the source resistance is very low (e.g., a
few hundred ohms or less) it is possible to drive the grid somewhat positive,
resulting in quite soft compression and very subtle clipping, although this mode of
operation is not easily achieved in practice. If the source resistance is very high (e.g.,
hundreds of kilo-ohms) then the voltage will drop more suddenly, causing harder
clipping and introducing more odd-order harmonics, and this is a common feature of
high gain amps."

- http://www.freewebs.com/valvewizard1...Gain_Stage.pdf

So what about that? Are you just speaking of power tube grids being difficult to drive in conduction, or preamps as well?

[loudthud]

If you are driving 6V6s and 6L6s at the same time form the same concertina, the 6V6s will clip before the 6L6s and tend to distort and clip the signal. Running more current in the concertina will limit this. A larger than normal grid stopper on the 6V6 will help also. I suggest you ditch the cathode bias for the concertina and use a voltage divider from B+ to ground like the Ampeg B25B. The bias point won't shift around with different tubes. You want the cathode voltage to be between 20% to 25% of the B+ voltage for maximum headroom. Install a temporary trim pot and tweek the bias.

[lowell]

Hi loudthud,
what do you mean the bias point won't shift around w/ different tubes? Can you explain this further?

[loudthud]

With cathode bias, the cathode voltage can move around quite a bit with different tubes. You might be tempted to stick a 12AT7 or AY7 in there. With fixed bias, (a voltage divider from B+ to ground,) the cathode voltage stays within a few volts unless the tube can't operate at the current the circuit demands.

[Merlinb]

" If the source resistance is very low (e.g., a
few hundred ohms or less) it is possible to drive the grid somewhat positive,
resulting in quite soft compression and very subtle clipping, although this mode of operation is not easily achieved in practice.

The grid stopper adds to the source impedance, so that it is no longer very low. Additionally, power valves usually need more grid current than preamp valves (although pentodes/tetrodes are not so hungry as triodes). And source impedance isn't the only issue- swing is too. That's why I put in my qualifying statement! If you make the loads really small you can get lots of quiecent current, but your voltage swing suffers instead, so you're no closer to getting what you want.

Ok, an ECC82 is a beefy preamp valve, and in the right situation it might drive some milliamps into a power valve, if a really small grid stopper were used. But with the concertina that will create a gain spike at the anode, which could cause some nasty intermod distortion, even if you acheive it.

Like I said, just try it. You're unlikely to damage the valve, and I think you'll discover that its neither as easy or tonally pleasing as you might expect.

[lowell]

Can you guys clear up the info on valve wizard concerning concertina load values? VW says to us 47k-100k for smaller output tubes (6V6), and to use larger values for larger tubes and more voltage swing. I don't get this. If the PI only provides .99% gain how can changing load values offer more gain? Either I'm missing something or the wording is not clear. Is it possible that what is meant by that is that the PI will PASS a larger signal applied to its grid if larger load values are used?

[Merlinb]

Is it possible that what is meant by that is that the PI will PASS a larger signal applied to its grid if larger load values are used?

That's exactly it. The load doesn't affect the gain, only the amount of output signal you can get before it clips.

If you have a small load then you might get, let's say, 50Vp-p from each output, and you need to put in 50Vp-p to get it. If you put in any more it will clip. You'll still have 50Vp-p output, but it will be a clipped output.

Increasing the load might get you, say, 100Vp-p output, provided you put in 100Vp-p. Any more and it will clip. And so on...