Question: about parallel and series (tubes/triodes)

[WickedLilMonkey]

Can you guys talk with me about the good and bad of running triodes in parallel and series
Why one over the other?
What can be expected from running 2 parallel triodes in series with 2 more parallel

Thanks for the help guys!

[soundguruman]

Can you guys talk with me about the good and bad of running triodes in parallel and series
Why one over the other?
What can be expected from running 2 parallel triodes in series with 2 more parallel

Thanks for the help guys!

What can be expected? Nothing.
The main application of using 2 triodes in parallel is to divide a signal into two paths. Or to mix two signals together.
For example I have two in parallel, one plate drives the clean channel, one drives the lead overdrive channel. Like in a Fender Supersonic amp.
Other than that expect no amazing advantage, there is not any.

[WickedLilMonkey]

See that's what I was thinking and have heard this from others but then there are those out there that say it'll give you a warmer tone and more touch sensitivity with a smoother overdrive... Marketing at its finest???

Edit: My guess is they are mixing the clean channels together and then driving the tube to an overdrive. So in essence you've got 2 equally overdriven channels mixed together. I guess?

[soundguruman]

Well if you need to MIX or SPLIT, it's just dandy and handy, but if you are just one path then you no need...

Series plate coupled stages allow building the gain, bit by bit, depending on the application of that added gain, cathode coupling gives you current for other special purposes, but no added gain.
But that does not mean that you can add together as many stages as you please...for infinite gain....
each stage adds noise too, and there's a limit to how much noise you can stand.

[rathmann]

Well if you need to MIX or SPLIT, it's just dandy and handy, but if you are just one path then you no need...

Parallel triodes will give you double the gain but half the noise of a single stage. It provides lower output impedance. Note that "double the gain" means very likely only a noticeable yet subtle increase in volume and "half the noise" means noise generated by the stage itself. It won't cancel noise from any previous stages nor hum from grounding issues. Overall there is some benefit from paralleled triodes, but it depends on what you want to accomplish. Typically if you have a "spare" triode in a design there are probably more interesting things you could do with it. If your design is such that you are looking for that "just barely overdriven" tone the parallel triode probably *will* give you a little more oomph especially if that stage is driving a tone stack due to the lower impedance. If you are the kind of person who feels strongly about the tonal differences of different brands of preamp tubes you might notice the different between a single triode and paralleled triodes. Otherwise the paralleled stage might just seem a little louder.

Series stages, as mentioned, are what you need to dramatically increase gain to where you can overdrive later preamp stages and the output stage. Two to three cascaded stages will typically give you anywhere from "classic rock" or "hard rock" tones depending on the design, where circuits using three to four gain stages are more likely to get you super high gain lead and metal tones. As soundguruman said, each subsequent stage will amplify the noise generated by the previous stage so flaws in design or lead dress can be exacerbated into a screeching banshee of squealing feedback or a Niagra Falls of roaring thermal noise.

A hybrid approach could reap some benefits. For example, a paralleled input stage will yield a hotter signal with lower noise and lower impedance than a single stage allowing more signal to reach subsequent cascaded stages especially if the first stage is driving a tone stack. Is this going to be an earthshattering difference than if you used only a single stage for the input? Not really.

I've used a series/parallel switch in a bunch of projects and like it quite a bit for having switchable gain. Could I simply switch out and bypass one of the stages? Sure. But I like knowing that nothing is going to waste and that the low-gain mode is a little bit better than it would be by simply bypassing a stage.

[Zozobra]

You can get some tonally interesting results with a parallel stage if you bias them differently.

[kg]

Parallel triodes will give you double the gain but half the noise of a single stage.

technically it provides double the transconductance. this may or may not net double the "gain."

[Chuck H]

Right. One of my eyebrows went up when I read "double the gain". I think that would only be the case if the two triodes were being used into a "load" for power rather than voltage gain. Which is never how I see parallel 12ax7 triodes used. OK, maybe as a reverb driver. In a typical preamp circuit I think the only significant tonal difference between a single triode and parallel triodes is the lower impedance. This reduces the reactance @ frequency (which increases with frequency). Since that usually means better HF I don't see how that's "warmer" either. I think it's just gear lore because Matchless did it a lot. But as noted, there are interesting and good things that can be done by paralleling triodes with different cathode and/or plate circuits. Like jumpering the channels in a four input Marshall. Series triodes is how voltage gain is ordinarily increased. When more triodes are in series than is needed to drive an amps power tubes the modern trend is to call it a cascaded preamp. One or two triodes in series is typically enough voltage gain for a guitar signal to drive the PI and power tubes to full power. But many high gain amps use three, four or even five "cascaded" triodes to heavily clip the waveform before it gets to the PI. The voltage gain does not just increase more and more with each stage though. There is a limit to what a tube can accept and produce. Outside of those parameters a tube clips the waveform and makes the distortion we all crave. So after a point, with cascaded triodes, we're just increasing saturation and clipping, not gain.

[riz]

I maybe can't explain it to you technically, off the top of my head, like some of the folks around here can (still working on that...), but in the one amp I paralleled the first stage (had an unused triode stock), it sounds better--just a little more of want I want to hear from an amp. It got a little "hotter" and "fuller" sounding... The way I did it shares the original plate resistor, with the two triode sections biased and bypassed a little differently, and I quite like the results.

[WickedLilMonkey]

I maybe can't explain it to you technically, off the top of my head, like some of the folks around here can (still working on that...), but in the one amp I paralleled the first stage (had an unused triode stock), it sounds better--just a little more of want I want to hear from an amp. It got a little "hotter" and "fuller" sounding... The way I did it shares the original plate resistor, with the two triode sections biased and bypassed a little differently, and I quite like the results.

This is a cool idea!


Great thread guys!

[exclamationmark]

There are some interesting things you can do, like sharing the plate resistor and using different bias on both tubes. Or you could even use different plate resistors and bias, and mix them together after the coupling cap (though there are some designs where you can do it before - lots of different ways to do it)

I'm pretty much oblivious to any tonal difference by paralleling a triode with the same operating conditions, though some other people could certainly appreciate the subtlety.

[J M Fahey]

Am I misreading this?
The OP talked about "parallel" Vs. "series" connected triodes.
In no moment he mentioned "cascaded stages" , *at all*, which is a very different thing.

there are those out there that say it'll give you a warmer tone and more touch sensitivity with a smoother overdrive...

Those are "magic words" which can be used to try to sell anything guitar related, from picks to output transformers.
Not necessarily mean anything in the real world.

Parallel triodes will give you double the gain but half the noise of a single stage. It provides lower output impedance.

Close enough.
In fact they provide twice the transconductance, as kg noted, which may mean higher gain *if the load resistor is kept the same*.
Problem is, since now we have twice the current through it, the static plate voltage goes down *a lot* , which may be quite bad , so usually the load resistor is halved.
Total gain thus becomes the same as before, what chuck h said.
It *does* lower the impedance, though.
The main advantage of paralleled gain stages is that the noise lowers by the square root of the amount of devices used.
So, with two triodes (or transistors, or Fets) you have the same gain but 0.707 the original noise.
To have an important noise reduction, you need a lot of parallel devices.
With tubes, it's an expensive way to lower noise but certain applications (funded by us all ) do not care about cost : Very sensitive long range Missile Defense Radars and Deep Space communications.
With transistors, it's cheap and easy. National Semiconductor had a very low noise IC whose differential input is made of 50 transistors in parallel each side.
Couple dollars cost.

[Chuck H]

JM, the only reason I mentioned "cascade" gain stages is because I percieved the Q as a "how to get more/better gain/distortion". So after my take on parallel triodes I thought a brief explaination of series triodes was in order to answer the OP's post.

[tubeswell]

Here are a couple of load lines for comparison

12AX7 triodes in parallel (you need to use your imagination to change the values on the vertical axis - by doubling the mA units)



Single 12AX7 triode



FWIW - in my quest for empirical, 'real-world' comparisons, I actually built both these circuits with identical 313V HT voltages and identical 110k load resistors. The parallel version had a measured plate voltage of 167, and the single triode had a measured plate voltage of 216 - Both had a bias voltage of 1.33V. (i.e.: slightly different from what the load lines show). What this does show is that the increase in current in the parallel stages' load resistor does decrease the idle voltage somewhat (as you would expect if the HT stays constant), but the gain with a 1Vp-p signal on the grid is about 75V in the case of the parallel triodes, whereas it is about 60V for the single triode. There you go.

[J M Fahey]

@ chuck h: I *know* you know
It's just that in my opinion the OP wrote the question with a wording that *to me* is a little confusing (although I have no doubt that he knows what he wants) so rather than guessing, I ask him to clarify.
@ mozwell :

I actually built both these circuits

COOL !! and very unusual. Way to go.
The only serious way in fact.
Loved that actual building and testing show what usually happens in the Real World:
1) *measured* results lie *very close* to *calculated* results.
Good and renews our confidence in calculations.
2) *very close* but not *exact*.
Why?
Among other things, because datasheets, curves, etc. are *averages* .
Real world commercial parts have production tolerances.
They usually behave quite close to datasheets (if the Company is serious) or all over the landscape if the Company is not.
Old tube amps (Blackface Twins as an example) had simple bias adjustments: "set bias to -52V" without need to scope them, add sensing cathode resistors, etc. , BECAUSE PARTS WERE CONSISTENT.
And often, Factory pre-selected.
Hiwatts and such, didn't even have adjustable bias ... but they used the best only.
Old Vox SS amps specified matched pairs of transistors, which were available from the Factory.
Today we need to add an individual bias preset for each power tube if we want them to really pass the same idle current.
Oh well.