PI tubes; balanced or unbalanced that's the question

[RockerTony]

Either the search engine doesn't work or this subject has never been discussed here....anyway, the question is:
Do you get balanced tubes for the PI? Does it make much difference? What's your experience?

[Alexo]

Nope, nope and none.

If you're building an 18, I presume you are doing it in an effort to recreate this classic, time-tested design. The original amps did not use balanced tubes, unless by chance, some wound up in them, so there is really no reason to seek them out for a recreation.

In guitar amps, the more "correct," or "technically superior" approach does not always correlate with what yields the best tone. The pi's are actually inherently imbalanced in these amps, so it would be a bit of a wasted effort anyway..

Having said that, if you think a balanced 12AX7 sounds better than an unbalanced one, then by all means, go for it!

[BooBird]

Tony, for bout tweenty yr's or so i balanced the PI for some of my better guitar player's, only the one's that could hear the sweet spot when we found it !
Here's what i learned about the PI, I could balance the PI by switching the tube, I had like 30K of preamp tube's to choose from, I would start out swapping tube's until i found the spot he liked, then i measured all voltage's and recorded those, what I found was it always ended up different on bout every amp we did this to ! and the voltage would always varie across the input of the PI tube, about 2.5 VDC to all the way out to 12VDC on some amps,
Thats the oldschool way of finding the sweetspot, you could do the same thing by changing resistor's, but a decade box would let you flip through some different resistors, I say if you can't hear these change's then it's not worth the effort,
If you put 1 probe of your DVM on 1 leg of the PI and the other probe on the other leg of the PI it'll show you the voltage imbalance between the 2 leg's,
I say it's like a lot of other stuff, to some it's the real deal, for other's it's a waste of time, i think it's up to you to decide if it's worth while !

[katopan]

Keep in mind that an 18 watt with 100k resistors on both PI plates causes unequal signal magnitude on each side anyway. Also the so called balanced PI with 100k on one side and 82k on the other (as different to the balanced valves you're talking about) still doesn't balance things and although it's better than the two 100k's the signals are still not equal, and to do this you would need a pot or odd resistor value combinations. The two common options above would seem to me to negate the effort of balanced valves.

[zaphod_phil]

And for balanced PI tubes to be effective, you would also need to have a perfectly balanced Output Transformer (they never are except in some hi-fi amps) and 100% exact value resistors and caps (they never are either). A slightly unbalanced PI and power stage will allow more even order harmonics to be produced, which add sweetness to the tone of an amp. Without them an amp will tend to sound somewhat sterile. So in guitar amps you don't even really need to have matched power tubes, and it may even be prefered for them not to be matched. Using carbon composition resistors for the 100k PI loads can also help produce some more even order harmonics to sweeten the amp's tone.

[BooBird]

I agree with you Phil, Matched PI tube's get you a matched tube !
But if you were to look at the balance of the PI on a scope and adjust that balance you'll see the even order harmonics come and go, if you look at the Dumble you'll see a trim pot added across the 2 legs of the PI, this was for adjusting the balance to get the even order harmonics that make the amp sound sweeter to your ear's, so if you check this with your volt meter, you'll see that there's alway's a certian amount of unbalance when your getting the good harmonics ! that's been my experiance anyway, take it for what it's worth, maybe 2 cent's,
Tone is in the ear of the Beer Holder !

[Alexo]

Using carbon composition resistors for the 100k PI loads can also help produce some more even order harmonics to sweeten the amp's tone.

If you have a cc resistor on each pi plate, and you generate equal amounts of 2nd order harmonics on each output of the pi, won't this 2nd order distortion be common mode and cancel out? Just like what happens in a push-pull output stage? Would this be more effective if just one of those resistors was a carbon comp?

[zaphod_phil]

I've wondered about that myself sometimes, and maybe one of these days I'll even have to get a 'scope out. However, the way I visualise it in my mind is more that a sine waveform coming out of one side of the PI has been slightly bent, due to the non-linearity effect. So the two sides of the power amp are amplifying slightly bent waveforms, which are combined in the OT and reach the speaker, just like any other waveform. I suppose what I'm saying is I don't think the CC resistor non-linearity is actually common mode.

[StarGeezers]

I ended up with 100k/90k in mine .... sounds Great!!! ... just Luck maybe !!!
Boo, what value trim pot in the Dumble "trick" , can't be much huh??? That sounds pretty easy...

[BooBird]

Geez, It's just a regular 10K trim pot, 2 outer leg's go to the leg's of the PI at the Junction of your .1 cap's and your 100k/90K resistors, the center leg gets tied to your B+, if you want i can send you a schematic of the amp i'm building, you'll need Nero photoshop viewer to look at it though, most of what i have is in Visio which i can't upload here !

[Alexo]

I've wondered about that myself sometimes, and maybe one of these days I'll even have to get a 'scope out. However, the way I visualise it in my mind is more that a sine waveform coming out of one side of the PI has been slightly bent, due to the non-linearity effect. So the two sides of the power amp are amplifying slightly bent waveforms, which are combined in the OT and reach the speaker, just like any other waveform. I suppose what I'm saying is I don't think the CC resistor non-linearity is actually common mode.

My head starts spinning when I try to visualize distortion characteristics in a balanced stage and then recombining, but...

Assuming all other things being equal (which will never happen) any distortion applied to both "halves" of the pi signals will by nature be applied symmetrically to both the positive and negative peaks of the signal once it's recombined.

Second order harmonics are by definition asymmetrical, so symmetrically applied even order harmonics would end up as odd-order harmonics once all is said and done - right?

Again, this is assuming the cc's are yielding an identical distortion to both pi outputs, which I don't think is really going to happen, so some asymmetrical distortion will be bound to creep through, but if only one Ra of the pi was generating distortion, it seems like that might maximize the even order harmonics you could get out of that stage.

...yep, a good ol' reality check on a scope would be nice, but I'd have to fix my scope first.

[Revv23]

That would assume that the signal is coming out of the CC exactly the same on both sides wouldn't it?

if the PI isn't balanced then I would imagine the harmonics produced by the CC's would be different as well, thus non-cancelling.

I'm pretty much a total n00b though so please shoot me if I'm off base.

[RockerTony]

Revv23, I might be able to understand your post better if I knew what CC stands for. Maybe I'm the noob here?

[Revv23]

We are talking that the even order harmonics that carbon composition (CC) resistors can add to the signal in the PI.

[zaphod_phil]

That would assume that the signal is coming out of the CC exactly the same on both sides wouldn't it?

if the PI isn't balanced then I would imagine the harmonics produced by the CC's would be different as well, thus non-cancelling.

Correct!

[RockerTony]

I should have been following this post better, after all I started it.
I've been doing my builds with mostly carbon film resistors. I understand that CC's have more shot noise than the other types CF and MF, but I didn't realize that this shot noise could actually add more even order harmonics. I'm going to try using CC's for the plate resistors of the PI tubes in my next build and see if I notice a difference. And BTW, no more balanced PI tubes for me!

[zaphod_phil]

Or maybe take Alexo's suggestion, and only have one of the resistors carbon comp, and keep the other one carbon film. That should help maximise the sweetening effect getting through.

BTW it's not the shot noise that's increasing the even harmonics. That shot noise is nothing but a nuisance. Rather, it's the asymmetric non-linear behaviour of CCs under high voltage conditions which provides the tone benefit.

http://www.geofex.com/Article_Folders/c ... oncomp.htm

[andrekp]

You want to keep the carbon films, metal films, whatever everywhere with one exception for using carbon comps: In an 18 watt, but the carbon comps in the PI plate load positions. (generally, in any amp, you want the CC to be in the driver's plate - only). There was an article somewhere that talked about they where and why - maybe someone has a link - but that was the long and short of it.

Also, I find these balance/unbalance conversations to get real confused because often people are not really talking about the same thing.

There are two kinds of balance in the PI: In the plate loads, and in the dual triode itself.

Sometimes people are actually asking about whether to $pring for a balanced 12AX7 for use as the PI (I thought that was what the OP was asking here at first, but then...) The general consensus seems to be don't waste your money because even if that tube is balanced, there are so many other things that need to be balanced to get the benefit that you'll just be bopping ferrets all day, as Gil says.

The other kind of balance, plate load balance, is further complicated by the fact that it is actually opposite initial assumptions. Lots of people confuse the idea of having 100K/100K plate loads in the PI with having the PI balanced. IN FACT, for technical reasons beyond immediate import, this is the imbalanced PI load. The balanced PI load will be someting like 82-92K/100K (YMMV).

I know this confused me at first until I figured out the decoder key. (and if I'm still wrong, someone correct me)

So, if you want an imbalanced PI for what some prefer as the best sounding amp, then use the 100K/100K CC's in the PI plate for an imbalanced PI. If you want as much balance as possible, you can either guess at 82K/100K, which is likely somewhat balanced, or tweek around and find the exact combination that suits you.

That's what I understand and I have tried a few combinations of balance, imbalanced, CF, and CC and found that I much prefer 100K/100K CC in that position. I use CF resistors everywhere else (excepting the power supply, obviously).

[lovedat800]

I tried various resistors and found that it's not a matter of balance being better than imbalance or visa versa, it's a matter of whether it works well for you and the amp. To my ear, more imbalance and the tone gets softer/squishier. So in an amp thats rather hard or harsh sounding, imbalance seems like it would be a great idea, tho adjusting the plate R seems a easier way then trying a ton of tubes. A amp thats already smooth may not benefit from imbalance, it may actually make it sound/feel worse because too much squish/smoothness can lead to a inarticulate tone. I was using several things in the amp to get a really smooth squishy sound, and providing imbalance was just too much after a while. So i think it's just a matter of whether the amps tone needs it or not. Like treble for example....you wouldn't want to add treble to an amp thats already bright or visa versa.

[stoo]

Seems to me you aren't going to whip out your resistor collection every time you change a PI tube? so TRYING to balance the PI is an excersise in futility.
Stew