Experimenting around with a tweed era circuit

[mort]

Up to now I haven't done much experimenting with parts of a circuit. I have a 5C5 tweed pro circuit built up but wasn't extremely pleased with the sound. It sounded a little like it had a blanket over it so I decided to toy around with the inputs. Here's the 5C5 version that I started with and it has the input running straight through some signal caps before hitting the grids in V1 and V2 and also has the cathodes grounded.

http://www.thevintagesound.com/ffg/s..._5c5_schem.gif

So I ripped out the input caps and 5m resistors and installed this exact input set up from a 5E5 drawing, and also added the 25uf/470 cathode arrangement.

http://www.thevintagesound.com/ffg/s..._5e5_schem.gif

The changes opened up the sound in he amp to a large degree. Sounds more present and bright. At first I had attributed most of the change to the inputs going directly to the grids but when I jumpered the cathode cap to compare I discovered that that's where most of the improvement comes from. I haven't tried putting signal caps back in (leaving the cathode cap) to see how that sounds.

So I guess my question is why have series input caps like in the 5C5? And are there any advantages to having the tube cathode grounded? Seemed pretty lifeless to me with it grounded.

Another curiosity is with the phase inverter. The 5C5 uses the paraphase but a later version of the pro uses a split load phase inverter. Does this effect the behavior of the amp or is it just a different choice?

And lastly, I also put a 12ax7 in V2 so that the instrument channel would become a sort of "hot channel". With the 250k voltage divider feeding the plates I'm getting ~100vdc on the 12ax7 plates and ~55vdc on the 12ay7 plates. What should I expect to hear different if the voltage divider is reduced in order to bring up plate voltage? Also do the divider resistors have to be the same value? Like if I wanted to keep the 12ax7 at around 100v but wanted to bring the 12ay7 voltage up.. supply one tube with one value, balanced by whatever is feeding the other tube, is that how that would work?

Any input or general conversation on adjusting tweed amps is appreciated.

[Justin Thomas]

Hey Mort,

I can only address the "why" of the way the early Fenders did things. And basically, they (most of the early amp manufacturers) were using published circuits from the tube companies. Fender ripped the circuits from the RCA, Sylvania, etc. suggested applications. Kinda like we start out cloning a Fender 5C5 and then go, "huh, that's not quite what I wanted... think I'll change some stuff." The difference is, Fender didn't have 50 years of Loud Crunchy Guitar to use as precedent... luckily, we do!

Justin

[mozz]

What you had is called grid leak bias. I did the same thing with a 5c3, a 6sc7 with grid leak bias i didn't like the sound. I changed to a 2.5k cathode resistor and a cap, takes humbuckers much better now. Also i tweaked the paraphase resistor from 6.8k to 7.1, done on a scope to help clean it up.

[Chuck H]

Indeed! Grid leak bias has it's limitations. Especially when pedals that involve boosting preamps are involved. You can almost ignore the cap input, the lack of grid stoppers (in that location they don't actually affect current significantly). It has a lot more to do with centering the bias on the tubes for more linear operation.

JM2C

[Raybob]

So I ripped out the input caps and 5m resistors and installed this exact input set up from a 5E5 drawing, and also added the 25uf/470 cathode arrangement.

http://www.thevintagesound.com/ffg/s..._5e5_schem.gif

The changes opened up the sound in he amp to a large degree. Sounds more present and bright. At first I had attributed most of the change to the inputs going directly to the grids but when I jumpered the cathode cap to compare I discovered that that's where most of the improvement comes from. I haven't tried putting signal caps back in (leaving the cathode cap) to see how that sounds. ...

You changed the cathode but left the 250k plate resistors? Curious to know if going to the 100k plate resistors improves the sound even more...

[mort]

You changed the cathode but left the 250k plate resistors? Curious to know if going to the 100k plate resistors improves the sound even more...

I just tried the 100k resistors which brought the plate voltage on the 12ax7 from 100v to 178v and on the 12ay7 from 50v to 100v. For my tastes I prefer the lower plate voltages. If I play real lightly with the volume turned half way up this amp has a real warm, vibrant quality with a touch of power tube break up and then when you play harder it gets mean and nasty, mostly in a good way. With the higher plate voltages it has a sharper, kind of harsh quality even when played lightly.

[eschertron]

Maybe I missed it, but do you have separate cathode resistors for the 12AY7 and the 12AX7? If it's still a shared Rk, each time you swap out a tube type the change in the combined cathode current will affect the behavior of the 'other' tube. I'm not saying that's a bad thing, but it introduces factors out of your control.

[mort]

Maybe I missed it, but do you have separate cathode resistors for the 12AY7 and the 12AX7? If it's still a shared Rk, each time you swap out a tube type the change in the combined cathode current will affect the behavior of the 'other' tube. I'm not saying that's a bad thing, but it introduces factors out of your control.

Thanks for noting that.. it is a shared cathode resistor and cap right now. That gives me something else to toy with.

In my reading, and correct me if any of this is wrong, I see that lowering the resistor value biases the tube hotter for more distortion and raising it does the opposite, and lowering the bypass cap value cuts bass response and raising it adds bass. And am I correct in thinking that lowering the plate voltage would bias the tube hotter (voltage goes down / current goes up to draw the same amount of total power?)

[Chuck H]

Thanks for noting that.. it is a shared cathode resistor and cap right now. That gives me something else to toy with.

In my reading, and correct me if any of this is wrong, I see that lowering the resistor value biases the tube hotter for more distortion and raising it does the opposite, and lowering the bypass cap value cuts bass response and raising it adds bass. And am I correct in thinking that lowering the plate voltage would bias the tube hotter (voltage goes down / current goes up to draw the same amount of total power?)

It's not exactly like that. The cathode resistor sets the bias point of the tube. That can mean on center, more prone to saturation or more prone to cutoff. This is relative to the input signal. A center bias will have maximum headroom amplifying signals up to the cathode voltage. Since a hotter bias has a lower cathode voltage, the same stage biased hotter will saturate more than a center bias in the same context. But a colder bias can be a source of distortion, but it's cutoff distortion rather than saturation distortion BUT since the colder bias will have a higher cathode voltage the same stage won't reach saturation until that higher voltage is present at the grid. Cold clipping stages (as they are sometimes called) are seen in many amps including Soldano and Trainwreck models as well as most modern uber gain designs.

In other words, it's not as simple as hotter means more output and colder means less. In fact, we take great advantage of a cold from center bias for the added power of an AB power circuit. The tubes on either side are allowed to reach beyond full power on one side of the waveform while resting in cutoff during the other averaging out to a (hopefully) safe dissipation.

Changing the plate resistor does change the bias, but not a lot. There is more current with a lower value but that is more because of impedance rather than voltage. The voltage will be more affected by the current as set by the cathode resistor (EDIT: The load and the bias are actually working together, but I'm trying to keep this short). In fact there is usually MORE voltage gain when a higher value plate load is used. This is also due to impedance because less AC is dropped across the load and dumped into the power supply filter. As for more distortion, the higher plate load usually makes for more distortion, but not because THAT stage is distorting, but because there is more gain at the grid of a following stage.

[Raybob]

In my reading, and correct me if any of this is wrong, I see that lowering the resistor value biases the tube hotter for more distortion and raising it does the opposite, and lowering the bypass cap value cuts bass response and raising it adds bass. And am I correct in thinking that lowering the plate voltage would bias the tube hotter (voltage goes down / current goes up to draw the same amount of total power?)

I would think it's more like opposite.. look at the high plate voltages of the blackface Fender 'clean' with Super Reverb, Twin running 260v plate voltages.

[Chuck H]

Ah! Yep... I should have mentioned that above. Lower plate voltages are sometimes a byproduct of higher current, but low plate voltages do not increase current and neither do most methods for reducing plate voltage. But to address the situation directly, if you run a tube at a lower plate voltage then center bias (or a hotter bias) will be a smaller bias voltage. So in that case the tube will reach saturation with a lower signal voltage at the grid. That is, a lower plate voltage does not in itself denote operating conditions closer to saturation, a greater grid sensitivity or a hotter bias.

[mort]

ok so I split the cathodes from each other and to keep the variables from being too many, put a 12ax7 in where the 12ay7 was. So now it's two 12ax7's, one cathode with the 470 resistor and 25uF cap and the other 12ax7 with a 2.2k and 33uF cap.

The one with 2.2k / 33uF jumped to 186v plate voltage and the 470r / 25uF is running around 90v plate voltage. Both channels are giving a pretty good bit more distortion and the little bit of clean head room that I had is gone(breaks up at any volume level) and it gets loud very fast on the volume control at around 2 or 3 and only gets more distorted and slightly louder all the way up the dial. The 33uF channel does have less bass than the 25uF one.

So is center bias on a 12ax7 always gonna be with a 1500r resistor, or is the center something that you sort of have to find?


And btw I don't have any sort of goal in doing any of this other than just running through possibilities for learning's sake.

[Chuck H]

I'm surprised the 2.2k cathode channel has less headroom now. Unless that was previously a 12ay7 channel?

1500r (1.5k) isn't an automatic center bias. Center bias needs to be determined by the voltage, the load (plate resistor) and the specific tube characteristics (not all 12ax7's are the same, especially today, though they may have been closer to a standard in the golden age). In fact 1.5k is usually a little cold (even though it's the standard in many guitar amps). Center bias is most easily determined with a scope so you can see if the wave form is clipping on saturation or cutoff. That isn't to say that center bias is ideal! At least not for guitar amps, which are actually signal processors if you think about it. FWIW center bias (with typical guitar amp voltages between, say, 125Vp and 250Vp) will usually require between 680R and 1.2k (depending on the specific voltage and load). So... You can see that your 470R cathode resistor is really very hot.

The higher and lower voltages you're seeing are a result of current due to bias condition (but you probably guessed that part). The tube has it's own internal resistance so more current means a greater voltage drop across the tube itself. If you prefer the tone of one bias condition (cold with the 2.2k and hot with the 470R) that would be the direction to follow. It's possible 470R is a bit TOO hot and may be stressing the tube. 820R is still a lot hotter than 1.5k (or 2.2k) and is typically the lower value used in guitar amps for a single triode. So probably a "safer" choice sans having a scope to look at what's actually happening.

[mort]

Yes, I noticed that the 470r tube was also getting much warmer to the touch than the other.

And the 2.2k was in the spot that first got the 12ax7, for whatever that's worth. And this amp had very little headroom to begin with, but didn't get quite as loud so early.

[eschertron]

Yes, I noticed that the 470r tube was also getting much warmer to the touch than the other.

And the 2.2k was in the spot that first got the 12ax7, for whatever that's worth. And this amp had very little headroom to begin with, but didn't get quite as loud so early.

You can tame the gain in the 2.2k side by removing the cathode bypass cap, you know, just to see. And that opens up the possibility of putting in small bypass caps, from well under 1uF to larger, to see what effect they have on gain and tone. For about any reasonably set-up stage, a cap larger than 10uF will be "fully" bypassed, so going to 22u or 33u will have no further effect on increasing the bass response.

And of course with the simple tone stack in that beast, there's not much loss - which equates to overdriving the following stages. My guess is that the designers did not expect the amp to be dimed.