X-chassis. Flexible Amp Design Platform

[silverfox]

Funy Bias Readings

Schematic Link: http://music-electronics-forum.com/a...seinverter.gif

Output section: 6L6GC-Output.pdf

Something I can't understand.

When checking the bias readings on the 6L6GC tube pair I get 53 volts at the bias feed point: R73, 77.

However by the time I measure at the 1 ohm cathode resistors the bias is reading 11 milivolts on each tube. And yet the amp sounds okay. I don't think it is putting out full power though. Consider also the protection diodes are grounded at the same point as the cathode resistors ground point- Shouldn't matter to me. Calculations based on B+ (440) and milivolt values indicates only 23 watts. I'm aiming for 35 or so...

Shouldn't I be reading the 26 mv at the power tube pin 8 to ground through the 1 ohm resistor? Another possibility is the circuit in question was designed for an output transformer impedance of 4 to 2 ohms. The tap I'm using is 16. Perhaps I should swith taps??

As far as the amp sounding okay, well, mostly. At times, depending on how hard the power amp is driven by an input signal, it sounds like it is faintly cutting out. Perhaps blocking distortion but not sure. Must be driven fairly hard before this occurs.

Thanks for the suggestions,

Silverfox.

[Enzo]

I may not exactly understand your question so let me know...

First, the protection flyback diodes don;t enter into it. If they get real leaky, then you usually get more serious trouble.

When I see a stock Peavey 5150 I expect to see the tubes idling at 11-15ma, this would be fairly normal for a high power PV tube amp.

But perhaps we are not on the same page. A common rule of thumb guys use it 70%. That means they like to bias the tube so idle dissipation is 70% of the tube's rated power. A 6L6 tube is rated at 30 watts, so 70% means 21 watts. Now you report you are aiming for 35 watts. That is way more than the tube can handle and it would be glowing red hot if you did that. So I have to wonder if you might not be confusing idle DC dissipation with audio output? They are not the same.

I cannot put B+ of 440 together with current through the tube of 11ma and come up with 23 watts. If your B+ is 440v at the plate, then 11ma through the tube means 440 x .011 = 4.8 watts. Add up 4 tubes of that and get 19 watts. Or six to get 29 watt. Thus my dilemma.

Bias current doesn't care about speaker impedance.

If you are using the stock PV OT with six power tubes, then it expects a 2 or 4 ohm load. But you report you are using a 16 ohm tap. A transformer designed for use with a certain number of such and such power tubes will match to its rated loads when used in such and such a power amp. It won;t affect idle current though. So whatever OT you use, maych the speaker load to the tap setting.

[silverfox]

Out of Office

Thanks Enzo.

The math you indicate is the part I don't understand with the circuit in that: I get (-)53 volts at the junction of the 47K resistors R73, 77 but at the grid, at idle, only 11 mv- (ma bias) is present. And yet the amp sounds fine for the most part.

Regarding the 440 B+ and power output, for that I just assumed the bias value must be 53 ma. Here is how I got that. Bias voltage at resistor junction is -53 volts. So the bias current at the grid must be 26.5 ma per tube.

But when I measure the bias at the 1 ohm cathode resistors on the power tubes I'm only getting 11 ma. (reads mv on meter)

(Do the 47K resistors shown on the output section schematic, the ones connected to the grids, are they knocking the bias value down and I actually need a higher bias voltage from the bias circuit to start with???

When I get a chance in the next couple days I'll post some values on the schematic as measured. The output section schematic is representative of as built. I used different Iron and reduced the tube count to (2) so 11 ma at idle is what is on each of the cathode resistors. Those resistors are added and won't show up on the schematic.

The flyback diodes are new so I don't think they are leaking. If they were I'd expect to hear some sort of square wave sound and perhaps some smoke ect..


I'll post something in a day or so.

Thanks,

Silverfox.

[Enzo]

If you have -53 at R73,77, then it ought to be on pin 5 of each power tube. (assuming 6L6) If you really had only 11mv of bias voltage, your tubes would all eb running flat out for all they were worth and red plating up a storm. But you report only 11ma through the tube (11mv across the series resistor of 1 ohm) For 53ma to flow through R73 or R77, there would be a voltage drop across each one of close to 2500v. That's Ohm's Law. If you tied the ends of R73,77 to ground, 53v dead across them, they each would draw: I = V/R = 53/47000 = 1.1ma


I think you may be confusing bias voltage and idle current. We set a voltage at the tube grid to bias it - to giver it a starting or idle current. The bias voltage is not numerically related to the tube current. Tube current is what you are measuring by its voltage drop across the 1 ohm cathode resistor. The way we set this up is to determine a desired tube dissipation at idle. We then calculate the current we need through the tube with a known B+ voltage present to make that wattage. Volts times amps is watts. Example: 500v across the tube and I want 21 watts dissipation. P = VxA. So, A = P/V = 21/500 = 0.042A = 42ma.

THEN I set the grid voltage - the bias - to whatever it takes to get my 42ma of cathode current. That could be 53v or 57v or 48v or some other amount.

[silverfox]

Design Descrepancies 6L6 Grid and Plate Resistors

I've really got to track down what is going on here and I think it has to do with the Peavey Sch..

If someone could look at these three comparisons of the as built and an ab763 schematic:



The Peavey circuit used in this project has (47K) resistors in the phase inverter at R73, 77. On the output tube grids there are 47K resistors. The screen resistors R88, 94 are (100R).

Compared with the Fender ab763 design which the Peavey is based in part on, the values are: 220K and 1.5K for the grids. Screen resistor values are 470R.

The bias values are different also with the Fender circuit running atat (-48) and the Peavey running (-58) as I recall.

The voltages I've got on the phase inverter and output section match the Fender schematic.
Should I change the resistors in question to the values used in the Fender output or is this just another means to arrive at the same destination regarding the power output?

To refresh the build history, the Peavey schematic was for a Roadmaster. I stripped the chassis and used lower power transformers from a different amp and also reduced the output tube count to (2), in an effort to build a 40 watt power amp version of the original Peavey. While it sounds okay there are points while I'm playing when the amp seems to be blanking or cutting out; but not completely. Sounds like two signals canceling or something like blocking distortion. I've tried to resolve a bias issue where the grids at idle read 11 mv but that can't be possible or the output tubes would be redplating. Crazy stuff I guess.

Thanks for the assistance,

Silverfox.

[jazbo8]

Not sure what is causing the oscillation problem, it seems to be some sort of postive feedback though... anyway, I don't think you need the 12AT7 cathode follower to drive just two 6L6s (or whichever tube you like to try), the Roadmaster only had it in there so it can drive 6 output tubes without stressing the 12AX7 LTP. So I think you can bypass it and just take the outputs from the LTP directly into the output tubes, while that simplifies the circuit a bit, I am not sure it will get rid of the oscillation - have you chopstick around to see if you can isolate the problem parts or wires?

[silverfox]

Back to the drawing board.

"I don't think you need the 12AT7 cathode follower to drive just two 6L6s"

That is not what I wanted to hear. It means the design employed was simply taken from a schematic without any in depth knowledge of the actual workings... Hmmm. I've seen quite a few of the pi circuits and don't recall seeing the two tubes together in the case of driving 40 or even 100 watts of output tube so you could be right. The blocking distortion may be another misdiagnosis on my part. The layout uses shielded cable to drive the grids and there are no long leads. The cutting out may be more related to overdriving the output tubes to saturation if the pi is stronger than needed.

At least it has worked sufficiently to allow some decent jamming on. Unless there are some very simple reworks to remove the AT7 tube it is back to the drawing board and since this is close to the ab763 perhaps I'll just mod it to that.

Thanks jaz.

Silverfox.

[jazbo8]

The cutting out may be more related to overdriving the output tubes to saturation if the pi is stronger than needed.

At least it has worked sufficiently to allow some decent jamming on. Unless there are some very simple reworks to remove the AT7 tube it is back to the drawing board and since this is close to the ab763 perhaps I'll just mod it to that.

Well, I wasn't really suggesting that the the CF was the source of the problem - it's a bit of an overkill for 2 tubes, but it shouldn't cause the the amp to cut out since it merely buffered the signal from the LTP - unless the tube was bad of course. You still need to address the oscillating issue which is likely due to a bad and/or in-correct connection. If you choose to take out the CF, you can simply swap out the 12AX7 with the 12AT7 and re-connect the grid stoppers of the power tubes to the coupling caps of the LTP, which should be pretty easy to do (at least on paper) - then you will have the AB763 configuration.