Mesa DC3 running cold with JJ EL844's

[booj]

Well it would figure I guess. EL844's are lower power and lower current tubes. The crossover notch is huge. I told the guy I'd make the bias adjustable, but I'm finding that is not so easy on this amp. Since there is no grounding resistor used in the bias circuit (page 3), a series resistor between the 33k and the two 330k resistors going to the grids has little effect. At the junction of the 33k and the 15k again referencing the bias circuit on page 3), I get about -15v, which leads me to believe the +2.3volts at the cathodes of the phase inverter (page 6) is basically dependent on the PI tube itself. That leaves changing the 120k resistor to a higher value. Does anybody out there have any experience with this type of biasing arrangement that could throw me a clue as to a good way to go with this? I have a feeling changing the 120k will do it but I'd like some feedback if possible. Thank you very much in advance.
boogie_dual_caliber_dc3.pdf

[Enzo]

Yessir, promise it first, then see if it can be done.

I suspect you are right, changing the 120k will alter the boas. it would also alter the PI cathode as well then.

Just me thinking here, I may not be accurate, be warned...

There is no filter cap at the rectifier for the bias, only at the other end of the 120k. Still we ought to have some idea the DC voltage there. And what is the voltage across that 220uf cap? The way it looks to me is you have some negative voltage at the rectifier, and the current through the PI tube sets the 2.3v. That makes the 120k and the 15k a voltage divider with the tap out the 33k for the bias. There is no current through that bias branch, so the voltage we want, in their case -12, is what appears at the cap. We don;t want to upset or change the PI, so as I see it in my admitedly fuzzy view, we have to move the bias tap along the voltage divider. Seems to me we want then to keep the 120k + 15k, 135k, between the rectifier and the PI cathode. After all, in many cases all that matters in a voltage divider is the ration, but in this one the tube current sets the voltage. I am assuming also you want to heat up the tube, so a lower voltage at the bias, like -10 or -9 or something, yes?


How i might try it is to change the 15k to a pot, with the 33k connected to the wiper. That should result in a bias control sweep of from +2.3 to -12v. I don't normally want to take the bias directly off a pot wiper, but in this case I can;t have shifting resistance like I would wiring pots up as variable resistors. Also, I don't want the bias to get anywhere near +2.3, so Maybe make that 15k resistance a 5k pot and a 10k resistor in series, with the pot at the end by the cap. Now that wiper still carries the bias voltage out, but it only draws it from the 5k portion of that voltage divider. In fact a little experimentation might find that a 1k pot and a 14k resistor is enough range.

And you may also find that the 2.3v has room to change without a negative effect, I have no idea. My rough calculations has about half a milliamp on each side of the PI, so about 1ma total. You have about 14.3v dropped across the 15k, which also gives me right about 1ma. The same 1ma that flows through the tube.

And I may be totally offbase with this, or there may be a simpler way, and someone can point it out.

[Jazz P Bass]

Why not simply 'pull down' point "E" to ground through a potentiometer or a resistor.
You will have to figure out a 'tail' resistor if you use a pot, so as not to go too low with the bias.

[Chuck H]

I had a Subway Blues with a very similar PA and power supply. As I recall I raised the screen voltage by jumpering over half the series resistors in the rail between the plate and screen node and then cathode biased the amp using a zener across the cathode resistor to "fix" the bias. I lifted the "E" supply from the 330k bias resistors and grounded that junction. I also pulled to 2.2M resistors. That leaves the CCS PI intact. The higher screens tighten the tone and the lower plate to cathode voltage loosens it a little. The higher bias current at the lower plate to cathode voltage definitely sounds better. I'm not sure what mojo Mesa thought they were getting with their DynaWatt circuit but it's MHE that these amps benefit greatly from generous Muntz-ing.

EDIT: It's also my experience that the OT in my Subway sounded very good and seemed to hide the dreaded EL84 BWIZZ effectively. These amps run the Vp too high and that necessitates a cold bias. They also drive a excessive signal into the power tube grids. The result is crossover distortion. If you find the crossover distortion too audible or objectionable at all you can always try the "Paul Ruby" mod. Works like a charm for EL84 amps.

[Enzo]

If we pulled point E with a pot, it would ultimately alter what appears at point D.

[booj]

Enzo, thanks for the reply. replacing the 15k with a pot could work since it wouldn't change the circuit in the PI's view. what messes with my head is that somehow the PI gets a positive voltage from a negative voltage. the other side of that 15k has -15 volts on it. Still the idea shows promise. another idea would be to have the guy put EL84's in the amp. Of course, the tubes would have to match the amp. smh.
Jazz P Bass (I have one of those by the way), I refer you to Enzo's second post.
chuck H, In my case I'd want to lower the value of the 2.2M resistors. Could work. And those resistors do offer a ground reference to the circuit. what is a CCS PI? What is Muntz-ing? What is the dreaded EL84 bwizz? But most importantly, what is the Paul Ruby Mod?

[Jazz P Bass]

Oops.

[52 Bill]

what messes with my head is that somehow the PI gets a positive voltage from a negative voltage. the other side of that 15k has -15 volts on it.

There should be a positive voltage at the cathodes of the driver tube that is being balanced with the negative bias supply voltage across the 15K resistor.

If you pull the driver tube, what happens to the voltage on the cathode side of the 15K resistor? For that matter what happens to the voltage on the supply side of the 15K?

[Mick Bailey]

With the PI removed there's practically no load on the circuit and the 220uf/63v cap would gradually charge to -330v via the 120k resistor. Maybe it would go pop first.

[Enzo]

It doesn't get a positive voltage from a negative supply. Remember, it is always about the circuit not the parts. You have a +300v supply and a who knows what negative supply. Strung between them is a circuit made of a couple odd plate resistors, and triodes in parallel, then the stuff we have been discussing. All along that circuit, the voltage moves from +300 to more and more negative voltages. At the cathode, it has not quite made it to negatives, but by the cap it has. That +2.3 did not just appear there, it is just a point along a continuum.

[Mick Bailey]

You have a +300v supply and a who knows what negative supply.

It's fed from 330vAC. The point I was making is without the PI there's no balancing in the circuit and the cap would charge to the peak negative AC value. So closer to 460v unless I'm missing something obvious.

[Enzo]

No argument at all with what you said, yes the cap would charge up to something high. That 330vAC translates into some DC voltage or other at the rectifier, just not sure what in this case. Maybe i am overthinking, but it doesn;t somehow seem appropriate to assume it will be -466, especially since there is no filter cap there. Was just trying to explain that the +2.3 doesn't just spring from a negative supply, but instead comes from a continuum between the high voltage extremes. The negative end needed a DC end point of some sort, otherwise i am trying to explain a continuum between +300 and 330vAC.