Crazy idea - full tilt sound, lower volume.

[2deaf]

I may be getting the hang of attaching.  Here are the pictures for experiment 1A.

[PRR]

> how to attach them.

When posting in the full reply box (not a quick-reply), click "Attachments and other options" and it opens up to expose "Attach:" and "Browse...". This lets you explore your machine and find your image file.

Please review the post at the TOP of this board about image sizes. Post a 4,000 pixel wide pic of your puppy in the snow, and Doug is likely to be peeved about server space (and we others get annoyed about screen space and download times).

Ah.... you got that part figured while my puppy was in the snow (film at 11).

You can also post to a public image server. I suggest imgur.com -- I do NOT like ImageShack and usually can't see images posted there. Imgur can be used free, they give you an explicit link-code for forum posts.

[2deaf]

Second try.  Pictures for experiment 1A.

[2deaf]

One picture

[2deaf]

I don't know where all those pictures are going, but they don't show up.  My ID appears saying I posted, but there is no post.

[2deaf]

Now nothing posts

[2deaf]

Finally showed up.

[PRR]

> they don't show up

Show up fine here. See attached.

May be a browser cache (or catch-up) issue. Press F5 to force (most) browsers to fetch the latest data for the page.

[2deaf]

Here are the pictures for experiment 1B

[PRR]

OK, when over-driven (90V peak on 50V bias with 50K series with grid) I do get a strong tripling (60%); but 2nd harmonic is 0.1%. Same-as a plain amp over-driven. (No real amp will get 0.1% 2nd due to small unbalances; that is exact-same model in all four sockets plus rounding errors giving a non-zero result.)

FOURIER COMPONENTS OF TRANSIENT RESPONSE V(R_R1)

 HARMONIC   FREQUENCY   NORMALIZED
    NO         (HZ)     COMPONENT

     1     1.000E+03    1.000E+00   
     2     2.000E+03    9.945E-04
     3     3.000E+03    6.357E-01
     4     4.000E+03    2.291E-03
     5     5.000E+03    1.658E-02
     6     6.000E+03    7.228E-04
     7     7.000E+03    8.850E-02
TOTAL HARMONIC DISTORTION =   6.420311E+01 PERCENT (60%)

[2deaf]

Sorry about the multiple posts -- I was down to trial and error complicated by a lack of patience.

Experiment 1A didn't have the complete frequency doubling that I thought it would, but it had an overtone that rendered it useless even if you could tolerate the overall unpleasant sound.

Experiment 1B did produce the frequency doubling, but the sound wasn't all that bad. Could be a really heavy Octavia pedal.

Both experiments had an ultra-slow oscillation with a period of several seconds.

[2deaf]

OK, when over-driven (90V peak on 50V bias with 50K series with grid) I do get a strong tripling (60%); but 2nd harmonic is 0.1%. Same-as a plain amp over-driven. (No real amp will get 0.1% 2nd due to small unbalances; that is exact-same model in all four sockets plus rounding errors giving a non-zero result.)

FOURIER COMPONENTS OF TRANSIENT RESPONSE V(R_R1)

 HARMONIC   FREQUENCY   NORMALIZED
    NO         (HZ)     COMPONENT

     1     1.000E+03    1.000E+00   
     2     2.000E+03    9.945E-04
     3     3.000E+03    6.357E-01
     4     4.000E+03    2.291E-03
     5     5.000E+03    1.658E-02
     6     6.000E+03    7.228E-04
     7     7.000E+03    8.850E-02
TOTAL HARMONIC DISTORTION =   6.420311E+01 PERCENT (60%)

Huh?

[2deaf]

Here is a set of reference numbers for experiments 1A and 1B.

[jeff]

 
Isn't inefficency the point?
Take a 100W amp plug it into a attenuator to get a 5W cranked sound.
I was thinking instead of an attenuator:
two tubes making 50W and two making 45W out of phase for a total of 5W cranked sound.
If you could make phase cancelation work=attenuated volume with no attenuator.

[jeff]

2Deaf
 In your experiments are you using two tubes or four?

[2deaf]

The schematics are four-tube with some tube symbols deleted.  When a tube is deleted on the schematic, that means it was pulled on the amp.

[2deaf]

two tubes making 50W and two making 45W out of phase for a total of 5W cranked sound.

In experiment 1B, the two out-of-phase tubes on the same side are not canceling each other.  Instead they are generating a signal that is twice the frequency of the input.  If you go back to my alternate reality post near the start of this thread, you may be able to see a plausible explanation for this phenomenon. 

[jeff]

Ok thanks, that's what I thought with the missing symbols.
I see what you mean with the 2 tubes and the doubling.

Did you just use two tubes because the 4 tube idea won't work? I don't understand why your test only involved 2 tubes. 

What I meant by 50W canceled by 45W was two in phase tubes push pull canceled by two out of phase tubes push pull with a slightly smaller input signal. The idea involves 4 tubes.

[2deaf]

I don't understand why your test only involved 2 tubes.

Primarily because I can do it with only unsoldering one grid and the NFB and also because two tubes is plenty to show what happens.  Any scheme that involves out-of-phase tubes on the same side of a class AB amp is going to suffer from frequency doubling of some degree.  Even the common mode rejection (experiment 1A) resulted in some doubling.

What I meant by 50W canceled by 45W was two in phase tubes push pull canceled by two out of phase tubes push pull with a slightly smaller input signal. The idea involves 4 tubes.

I understand what you mean, but the frequency problem is the monkey wrench in your works.

Attached (I hope) is a picture of a 5881 signal with 16 Vp on its grid.  The signal never swings negative, it is either zero or some positive number.  Signals can still cancel if their combined signals always equals the same number because that is DC and DC will not induce a current in the secondary.

[jeff]

I guess I'm having understanding why you'd have any output at all in experiment 1A
 If you were using one tube and attached both ends to the two ends of a CT OT that should cancel, right.
 I can't understand why, if you're using two identical tubes, with identical bias, identical input singal, in identical phase with their plates attached to each end that wouldnt cancel too.


That's the part I don't get what's the difference between:
 
   tying both ends of an OT to one tube's plate
                           OR
   tying one end to one tube's plate and the other end to an exact copy of that tube(in every way)?

What am I missing?

[HotBluePlates]

I guess I'm having understanding why you'd have any output at all in experiment 1A
 If you were using one tube and attached both ends to the two ends of a CT OT that should cancel, right.
 I can't understand why, if you're using two identical tubes, with identical bias, identical input singal, in identical phase with their plates attached to each end that wouldnt cancel too.
...
What am I missing?

I pretty much dropped out of the thread a while ago, for one reason: the whole idea seems extraordinarily impractical for the goal you want to achieve, even if it did work. Because there are simpler (and cheaper, in terms of tubes, OT, power supply and parts) ways to do the exact same thing you want (phase-cancellation to reduce output volume).

In the originally-proposed form, it's as if you want a 10 horsepower motor, and chose to get there by using 2x cars with 300-hp engines, chain the back bumpers together, and push the gas pedal down slightly-less on one of them (to get the net effect of 10-hp pulling in one direction). For a mental exercise it's fine, but ...

[sluckey]

 

[2deaf]

with their plates attached to each end that wouldnt cancel too.

It does give the signal a -24 db whack which is some pretty serious cancellation.

to get the net effect of 10-hp pulling in one direction

Hey!  That's a cool idea!  Has anybody ever tried it?

[sluckey]

Won't work with just two cars. Gotta have 4! 

[HotBluePlates]

Hey!  That's a cool idea!  Has anybody ever tried it?

The problem is the steering wheels. If they're not locked in one spot (exactly on the same plane), the two cars rip off each other's bumpers.

Come to think of it, that's not a bad analogy for the cancellation of output tube outputs... There's bound to be some degree of inequality that will throw it out of whack.

[jeff]

Ok stupid idea
I feel I need to explain why I started thinking about this.
As far as cheaper ways to do this I agree. It's cheaper to build a 5W amp and crank it. I was thinking more along the lines if you already have a 100W amp and love the way it sounds but want it quieter it'd be cheaper to buy one switch and one pot than a speaker attenuator.
Would this be ineficent if it worked, for sure, but more or less inefficent than running a 100W amp through an attenuator to get 5W out.
To take the car analogy:
If the speed limit on your street is 25MPH you can drive your muscle car at 25MPH. But if the whole point is that you love the sound when the tach hits the red line maybe there's a way to get that sound and feel at 25MPH.
By going 250MPH while tied to a car going 225MPH in the opposite direction your still only going 25MPH but it's going to feel and sound much different than just driving 25MPH. Yes that is a costly, extreme and rediculious waste of power.

But when you really stop and think about it, isn't cranking a 100W amp with an attenuator to get that sound and feel at 5W also a costly, extreme and rediculious waste of power.

If phase cancelation is another car pulling against your car then I would say that an attanuator is like having the pedal to the metal with the emergency brake on. Neither one really make sence if the only point is to drive 25MPH, but if the point is to redline the engine at 25MPH....

Just felt the need to explian myself. So that was my idea. Cheap way to reduce he volume of your cranked 100W amp.
But by talking about it I was learning more about phase cancelation and OTs.

Thanks for putting up with me trying to think outside of the box. My crazy ideas don't usually work but I do learn something when you guys help me to understand why they won't.

Thanks
Jeff.

[John]

Thinking and wondering-why for yourself is never a bad idea. Accepting conventional wisdom is usually fine, but how do you learn?


On a sort of related thing, has anyone here played through one of those attenuators? I'm thinking that cranking it down very much probably takes away a lot of the feel of playing cranked, since (imho) so much of the feel comes from the speakers moving air. I've never used one, just curious.

[2deaf]

by talking about it I was learning more about phase cancelation and OTs.

I know they have grown weary of this thread, but as long as you are learning, I would like to point out one more characteristic of transformers which is of no concern in normal utilization but which manifests itself as a problem is this application.  The secondary can induce a signal in the primary.*  When the north side of the primary induces a signal in the secondary, one of the things that happens is that the secondary induces a signal in the south side primary.  The south side signal is inverted because the south winding is the opposite of the north.  In experiment 1A, the south side is already getting a non-inverted signal so now it has inverted and non-inverted signals at the same time which is the same thing going on in experiment 1B.  By amazing coincidence, the 16V picture from experiment 1A bears a remarkable resemblance to picture 10mV from experiment 1B.

*OT's can act like an automobile ignition coil if you hook the points up to the secondary and a spark plug up to the primary.  Apparently, somebody tried it with tubes instead of a spark plugs and we've been hooking diodes to the plates ever since.

[2deaf]

On a sort of related thing, has anyone here played through one of those attenuators? I'm thinking that cranking it down very much probably takes away a lot of the feel of playing cranked, since (imho) so much of the feel comes from the speakers moving air. I've never used one, just curious.

You might want to start a new thread on this subject.  I have a few things to say about it.

[HotBluePlates]

Ok stupid idea ... I was thinking more along the lines if you already have a 100W amp and love the way it sounds but want it quieter ...

No, not a stupid idea.

Free Engineering Lesson:
1. Start an investigation with a desired result, or end product (here, reduce output power of an existing 100w amp).
2. Brainstorm all possible approaches to achieve the desired result. As in ALL ("stupid" is encouraged at this stage, to allow for novel or innovative approaches).
3. Establish all factors needed to judge suitability (cost, weight, complexity, reliability, ease of manufacture/assembly, etc).
4. Prototype. Try out your ideas. See what works & what doesn't in actual practice.
5. Evaluate the prototype(s) against the factors you established earlier. You could do this before prototyping to save time & money by avoiding prototypes likely to fail, or which will be judged unsuitable by your criteria. Assign a relative weight to each factor to show how important it is overall. Score the "Design" against the criteria, then multiply by the weight for each factor. Add them up. The highest (or lowest, depending on your scoring scheme) scoring design will be the best compromise
6. Implement the design chosen to best meet the criteria, and proven workable after prototyping.

So "bad ideas" are allowed and encouraged. The only poor choice is to get married to one idea to the exclusion of all others at the concept stage.

I think many of the posts proposed alternative ways/places to accomplish the phase cancellation you thought of (rather than at the output tube plates), or raised possible problems which might "score poorly" against the suitability criteria. Of course, people build "bad ideas" all the time, maybe because they have on-hand parts and the cost of getting the right stuff (or some other factor) rules out "better plans".

[jojokeo]

A crossline MV isn't all that bad of a bargain...only a single pot to achieve what jeff wants for about $2.50 plus has a much better range without most of the negatives too?