12ax7 bias

[daz]

I have again been playing with preamp values in my amp which is a marshall style affair with a front end vaguely similar to a JCM 800, at least in how many cascaded stages it has. I recently decided to start gigging again after 6 years or so, and this is the first time i have regularly used this amp in a band situation. But at rehearsals i realized the tone was softer than i wanted so i lowered the values of thee dropping resistors i had in the power rail. I had been using very large ones to keep voltages low, and while i liked that, in the band it needed more articulation and the low voltages caused it to be too soft sounding. this really did the trick, but it also changed the tone in ways i needed to rectify because the rest of the amp values were chosen with the low supply voltages.

That said, I tried sever different values for the first two stages cathode resistors and was surprised to find that no matter what i tried, the best tone i got was with 820R on both stages. This seems odd to me because looking at other schematics of similar amps, no one seems to do that. Not understanding 12ax7 biasing and load lines and all that, i'm curious as to why this is. In layman's terms can someone explain why i get much better tone with those values? The tone has more complexity, is much more articulate and fuller. I assume it's because when i brought the supply voltages up it threw the tube's bias to the outer edge of thier range and lowering the cathode R's brought them back to where they should be. But why the better tione? Because it's now in a similar bias range but the tube is more dynamic with higher voltages? But then why do some amps use low voltages like the bad cat with 99v on the first 12ax7's plate?

[MWJB]

You don't say what current draw & late voltage you had before going down to 820R...not that it matters, if you like 820R at the voltage you are currently running, then that's what you like, so stick with it.

Be aware that by changing the cathode resistor you are changing more than just plate current (bias), you are significantly changing the tone too.

E.g. Let's say we have a typical Fender style triode fed by 350vdc off the B+ rail, with 100K plate R & 1.5K cathode R. Plate voltage will be around 235v, plate current will be about 1.15mA. If we wanted to try lower current, the most transparent way of doing this is to drop the voltage fed at the B+ rail by using a larger power supply resistor. This also obviously changes the plate voltage, but gives the more consistent tone. Dropping the plate voltage to 175v would mean reducing feed at B+ to around 260vdc. Plate current at this voltage would be in the region of 0.85mA.

Now bear with me because leaving the power supply alone (350vdc feed) and just subbing the cathode resistor from 1.5K to 820R would probably give the same final plate voltage (or at least ball park) but at higher plate current - 1.75mA at 175vdc on the plate, compared to 0.85mA.

In short what I am saying is that subbing cathode resistors has a big impact on tone & texture in the preamp, more so that at the power tubes over typical changes. The higher the preamp tube current (for a given plate voltage), the lower the headroom and crunchier the harmonics. Lower the current and you get more headroom/cleans.

What you have discovered is basically that you liked the "flavour/texture" of the 820R cathodes, but needed a little more B+ to realise better headroom & dynamics?

I doubt that you had the tube's biased to the outer limit at the higher voltage, cathode values of 10K (unbypassed) are not unheard of in stages where transparency is desired. The tube will probably work fine in excess of the recommended plate voltage. However, you can occasionally run into problems with going too high on the current, at low voltage, where the tube begins to get too muddy, possibly choking under heavy loads.

So if the tone is "nearly there", just needing a little nudge, adjust voltage at the power supply dropping resistor. If you need to significantly change the character, then tweak the cathode resistor. Normally a designer will stick to a given plate/cathode resistor value combination to give their trademark character (e.g. Fender 100K/1.5K per triode) and tweak B+ to fine tune response. I guess I'm saying that 820R at 150vdc plate will have more in common, tonally, with 820R at 200v on the plate, than with 1.5K at either of those voltages and then some at either side.

[daz]

Thanks. i will need to read that a few times to allow it to sink in, but i got a lot of it. I can tell you this....I played around with the supply after i discovered this and the 820's didn't sound as good with the supply lower. it seems like this is the place to be.

One thing that hit me and i will try it tonite when i get home is this...after i upped the supply voltage i tried several cathode values and anything larger than 820 wasn't as good. So even tho 820R is the smallest i can recall seeing, i'm gonna try a 470R tonite. I mean hell, it now seems the lower i go the better it sounds/feels, so i may as well keep going and see if it improves even more till i hit a wall on the low side. That may be my one advantage over you guys....i'm too stupid when it comes to design to know not to try something too unorthodox but for some unexplainable reason may actually work well.

[sean k]

I remember this guy telling me years ago about some chaps he knew who ran a set of power tubes at about 40% above the max voltage ratings, I didn't know enough then to ask whether they were that much above the max power ratings, and the tubes worked beautifully... but their useful life went down expotentially to the point where they lasted a matter of hours as opposed to a few hundred hours. Things do look best when they burn the brightest... but they don't last anywhere near as long.

At 470 ohms you're almost doubling the anode dissipation while pushing the bias much closer to the 0V point.
I know from what you said you don't do load lines but if you're changing out resistors in the powersupply and suchlike you've most probably got a multimeter and I'm sure you can do a little math. Take a voltage reading at the cathode as it is now and take another at the plate. Divide the cathode voltage by the resistor value you have and note this figure. Now times that figure by the plate voltage and you have the wattage at the plate.

That final figure wants to be under 1.2 and if you change out the 820 for a 470 that figure will almost double. Before I knew all this stuff I put a switch on a cathode to ground it! didn't actually change the sound that much but I know know it was definitely shortening the life of the tube.

The Valve Wizard has a downloadable section on 12AX7's as preamp tubes and he explains all the in and outs of raising and lower values in and around the tube and from my reading the plate resitor is also very tunable for what you want. Lowering or raising that , by increments, possibly may do more for you than shifting the bias point left or right.

Plate load first then cathodes then powersupply I would think might be the way to fine tune where you want to go.

So I've read his stuff, the Valve Wizards, and I've built one or two amps that still work, but I'm still a complete newbie shuffling about mainly in the dark, but as far as I can see the plate resistor sets your region of harmonic content, the cathode resistor defines the way you want things to distort and the powersupply defines the headroom you have... all the while keeping an eye on the plate dissipation or wattage on the plate defined by the current in the cathode resistor and the voltage on the plate.

[daz]

That badcat i spoke of actually has no cathode resistor on the first stage, it grounded. It's got a 220k plate R and 99v at the plate. Just for the record. I don't understand much as i said, but somehow i guess from what you said and looking at that, the lower the supply V the lower the value of the cathode can be w/o running it too hot?

[sean k]

Basically yes. Ohms law isn't too much of a hurdle to get over and a lot of stuff opens up once you get it. Just a multimeter to chock DC voltage and the odd checking of resistors and you can basically play doctor on tube amps with abandon.

[tedmich]

Merlin clarifies this greatly here:
http://www.freewebs.com/valvewizard1...Gain_Stage.pdf

or buy his whole book for $38 at Doberman amps!

PS 820 is about as low as you can go on cathode resistors, highest gain, no bypass caps though eh? That would likely be too much (blocking dist. here I come!)

[daz]

Actually V1A is not bypassed and the amp gets no blocking distortion at all that i can tell even with the 820R. But on your suggestion i'll keep that in mind and maybe try bypassing it with a cap and resistor in series to minimize the tonal effect because it already has as much gain as i want, and a 22uf cap to minimize tonal changes would add too much gain. Or is there any other way to do it w/o adding much gain or tonal change?

[mooreamps]

Merlin clarifies this greatly here:
http://www.freewebs.com/valvewizard1...Gain_Stage.pdf



PS 820 is about as low as you can go on cathode resistors, highest gain, no bypass caps though eh? That would likely be too much (blocking dist. here I come!)

Well, I don't think his graphic on grid leak is good enough. He misses the point where for grid leak to work, requires an in-line cap to capture the grid leak current. It's not shown here, thus fails to illustrate the point. Therefore, if set-up correctly would provide about a 1/2 a volt of bias with the cathode connected directly to ground. Also, it's not presented as clear why having the 10 meg input impedance on a grid leak stage makes it more sensitive to noise. Not something I would use for a guitar amp, personally. I once thought of using grid leak for a pizo or a hi-z mic input, but choose a different type of input instead.

-g


Further, I have yet to find an of the old hi-fi schematics that used fixed bias.