Phase inverter distortion

[rifraf50]

Hi, having trouble with the PI in a build I am currently working on......I am getting lots of uneven distortion and I just cant seem to get a handle on it.
The strings on my guitar seem to be VERY touch sensitive and distortion is at all volumes

I have redressed it 3 times thinking that maybe this or that wire is to close etc but every time ends in the same distortion.
There is no hum or audible oscillation except when you go to 90% on the master....then a squeal occurs

when I sub the 12ax7 with a 12au7 it tames to almost bearable audio quality...just a very small amount of said distortion and no squeal

I changed the power cap to be sure and no change....

The wiring seems perfect to me... each time making an effort to separate the wires a little more but to no effect....iv'e tried different tubes to no effect,changed caps to no effect and resistor values to no effect.

I even bypassed the PI and ran a cap to each output tube and.........NO distortion.....loud and clean
Then I gave up on the long tail PI and put a boot straped cathodyne PI and still it distorts with the same distortion....UUUGGGHHH!!!!

With only one triode handling the PI with no gain apart from the bootstaped triode I am now out of ideas

I tried the park le then the fender.....no change

I will post the schematics of the tmb and the fender...but for the life of me I cant see a mistake anywhere......Help!!!

 

[JonSnell Electronic]

Have you checked the wave form from the phase splitter? It is usual, as the ECC83/12AX7/12AU7/6SN7 etc are not balanced, gain wise, to adjust the anode load resistors accordingly. Pin 1 with 100k and pin 6 with 82k gets it somewhere near.

 

[Sorento]

I even bypassed the PI and ran a cap to each output tube and.........NO distortion.....loud and clean ...

how is this possible ?
w/o PI from where do you get complementary signals to drive the outputs ?
if that worked, what do you need a PI for in the 1st place ;-)

if that means that when you drive both outputs with the same (!) signal, you get undistorted sound, then you should think about that first ...

in that case it sounds like you got your wires on the OPT's primary crossed, something like one EL84 connected to the center tap and B+ (A) to one end of the primary ... ?

 

[JonSnell Electronic]

With no phase splitter it is running single ended and if both output valves are excited at the same time there will be no output. Check the drive on your oscilloscope, as I suggested, for phase shift and symmetry.

 

[rifraf50]

how is this possible ?
w/o PI from where do you get complementary signals to drive the outputs ?
if that worked, what do you need a PI for in the 1st place ;-)

if that means that when you drive both outputs with the same (!) signal, you get undistorted sound, then you should think about that first ...

in that case it sounds like you got your wires on the OPT's primary crossed, something like one EL84 connected to the center tap and B+ (A) to one end of the primary ... ?

hi,it was just a test to see if distortion was running through the output stage....I didn't run both tubes but one at a time to check all is ok without using PI...

 

[rifraf50]

Hi,i did balance the long tail PI with a 82k resistor but no change.....I tried the park le cathodyne PI and then went to the fender design going from two 100k resistors to two 56k resistors......no change

I figured the cathodyne PI would be more balanced as it uses only one triode...an unbalanced long tail is more probable if the tubes triodes are not matched

long tail gives you gain but cathodyne does not which is why I attached the extra triode to it (boot strapped)....in the end you end up with more gain than long tail PI but more chance of instability which is why the grid stopper is 470K as in the park le schematic

still stumped !!!

 

[rifraf50]

oscilloscope ?

Yeh,I cant check that part (wave form) as I am in the process of acquiring a scope as we chat...but it seem conclusive that that something is turning it south...

Ive built this one before with no issues , nice amps....strong and pedal friendly....but never had this problem before, this is why I'm calling for help , can't seem to find an answer anywhere...

 

[famousmockingbird]

You show three schematics which one did you build?

One has feedback so maybe you need to reverse leads to make it negative and not positive. This could explain why when "bypassing" the PI you don't get the issue, the feedback is injected into the PI.

 

[Sorento]

In the OP you wrote that you have distortion with any PI whatsoever;
and distortion goes away when single-ended;

doesn't that point to a problem with the outputs and not the PI ?

This is why I proposed to check whether OPT primary is wired correctly.

If you inadvertently connected plate of one output tube to the center tap of OPT primary, and power supply to one end (!) of primary then you get massive amounts of distortion with any PI in PP but not in single ended with just one tube active ...

 

[Gnobuddy]

Pin 1 with 100k and pin 6 with 82k gets it somewhere near (balance).

This is true only if the "long tail" is, in fact, not very "long". If the tail resistor is greater than about 20k, then using 100k on both anodes will be closer to balance than the old 82k/100k combination.

There is also the question of whether a perfectly balanced phase splitter is a good thing or not - a few percent imbalance in either the phase splitter, or output stage, sounds better to many guitarists. (A hypothetical perfectly balanced push-pull stage completely cancels all even harmonics; introducing some imbalance here allows some even harmonics to make it through to the speaker.)

-Gnobuddy

 

[rifraf50]

wires are correct,

In the OP you wrote that you have distortion with any PI whatsoever;
and distortion goes away when single-ended;

doesn't that point to a problem with the outputs and not the PI ?

This is why I proposed to check whether OPT primary is wired correctly.

If you inadvertently connected plate of one output tube to the center tap of OPT primary, and power supply to one end (!) of primary then you get massive amounts of distortion with any PI in PP but not in single ended with just one tube active ...

I have another amp in which I have a lead plugged in and a capacitor attached to the core and the shield attached to the chassis......with this I amp able to hear what is coming out of each stage and everything is fine until it comes out of the PI

The OT is a fergusson OT of Australian make and is a high fidelity make 20htz/20KHTZ and is multitaped for either 6K or 8K operation (pp).....its currently setup for 8K operation.....it is wired the correct way and I have swaped the plate leads without effect (colours and ohms tell all).The center is not perfectly centered but I believe this wont cause an issue...so the saga continues....

 

[rifraf50]

You show three schematics which one did you build?

One has feedback so maybe you need to reverse leads to make it negative and not positive. This could explain why when "bypassing" the PI you don't get the issue, the feedback is injected into the PI.

hi,started with the long tail then the park le and ended with the fender style.....just cant get my head around something so simple is not working...is there some kind of inaudible parasitic oscillation happening here ?

volts are slightly higher by 15 - 20v on a given day,les than 10% though

 

[rifraf50]

distorts on all 3

You show three schematics which one did you build?

One has feedback so maybe you need to reverse leads to make it negative and not positive. This could explain why when "bypassing" the PI you don't get the issue, the feedback is injected into the PI.

Distorts on all 3 so at the end I went fender as it looks a little less dynamic....ugh !!

 

[rifraf50]

I have another amp in which I have a lead plugged in and a capacitor attached to the core and the shield attached to the chassis......with this I amp able to hear what is coming out of each stage and everything is fine until it comes out of the PI

The OT is a fergusson OT of Australian make (OP308 8/4) and is a high fidelity make 40htz/30KHTZ and is multitaped for either 6K or 8K operation (pp).....its currently setup for 8K operation.....it is wired the correct way and I have swaped the plate leads without effect (colours and ohms tell all).The center is not perfectly centered but I believe this wont cause an issue...so the saga continues....

"note the OP change...40 / 30khtz"

 

[rifraf50]

is the tranny

Hi guys...something strange happened today as I once again tried to make this amp work...

It seems that the PI was creating some distortion ....I came to this conclusion when I jumped the PI and ran 1 tube at a time...

Each tube ran perfectly in class A ,so I concluded there must be something wrong with the PI......WRONG....

I pulled a tube on each side and discovered that if I pulled a tube,one tube would run in class A through the PI perfectly and yet putting 2 tubes in would result in crazy distortion

I know that the windings are not well centered as the tranny is geared for 6K or 8K operation.....from B+ one reads 119 ohms the other reads 97 and 133 ohms

Ive also noticed that as I power it up it will run clean until its up to operating temp and then becomes totally unusable

Being relatively new to tube theory I have not encountered such a problem yet

I have also rebuilt the PI back to the original long tail version and still a 12au7 will tame to be almost usable but without the volume I want.

It seem that the transformer is becoming unstable with the extra current but even at low volumes its still distorts (with a 12ax7 in place)

Surely if blocking distortion was an issue it would show up when running just a single tube ?

its driving kinda nuts now........HELP!!!

 

[Gnobuddy]

I sympathize with your frustration. You've already put a lot of work into trying to fix this.

Could it be you have a defective tube? Gassy, perhaps, so that thermal runaway occurs as it heats up? I've never encountered this, but Tubelab (George) has mentioned this type of runaway in several of his posts. He has also mentioned tubes that exhibit runaway - after a good long period of apparently stable operation - if you put too much voltage on their screen grids.

At some point, the simplest fault-finding method is worth trying: swapping in known-good components.

-Gnobuddy

 

[rifraf50]

All in the quest for knowledge !!!

Thanx for the reply....I was kinda thinking there the thread had gone stale

Well the tubes are stable...ive tried dutch , Hungarian , jap and JJ's and they all exibit the same problem.......so tubes seem to pass the test

As for the running clean until operating temp I'm thinking because amp is not at 100% its running as if its using lower gain tubes...same as subbing the 12ax7 for the 12au7 in the PI

Ive been reading a little about hysteresis in output tranny's , a condition where the core becomes magnetic and causes distortion at all frequencies.....is this indicating too much DC current running through the windings ?

Ive got another OT I can A/B with to see what effect it will have...that seem the next logical step.

I'm curious about hysteresis now....can anyone explain it better than some of the truly technical scripts I've found...(dumb it down for me)

 

[Gnobuddy]

...so tubes seem to pass the test

To make sure I understand: the amp works perfectly with many different tubes, but in every case, once it's reached full operating temperature, it distorts like crazy? Is that right?

If so, something is responding to changing temperature. Are there any components mounted very close to the tubes that might be overheating? How about the tube sockets, are the sockets making a tight grip on all pins? All solder joints on the sockets wiggle-test good? No resistors with drum-tight leads that might actually pull away when heated?

As for the running clean until operating temp I'm thinking because amp is not at 100% its running as if its using lower gain tubes...same as subbing the 12ax7 for the 12au7 in the PI

So is the amp is well behaved at all temperatures as long as its not set to full gain? If this is the case, you almost certainly do have instability, probably oscillation at frequencies too high to hear.

What does not make sense to me is that you are apparently having this problem regardless of the actual circuit topology of your amp - you've tried multiple different configurations ahead of the output valve, and all of them misbehave, is that right?

I am still not sure exactly what symptoms you're dealing with - it's very important to pin that down exactly first, so that you can then make a logical hypothesis as to what is wrong, and then test that hypothesis by an appropriate experiment. In this way, you play "whack-a-mole" with each hypothesis in turn, until you find the correct one, and fixing that will cure your amp. There is no black magic necessary, I assure you, only cold, hard, persistent, stubborn logic.

Ive been reading a little about hysteresis in output tranny's

Did you try swapping in a different output transformer to see what happens? You have a "mystery problem", and when a problem defies diagnosis, one way out of the impasse is to keep swapping components until the problem goes away.

I've got another OT I can A/B with to see what effect it will have...that seem the next logical step.

Agreed! Please do try that!

I'm curious about hysteresis now....can anyone explain it better than some of the truly technical scripts I've found...

In a nutshell, a perfect, linear, transformer would accurately reproduce (on the secondary winding) whatever signal you feed into the primary side. This happens because the core responds to the primary signal by magnetizing itself in an exactly matching way.

In reality, the magnetic material in the core has imperfections that prevent this from happening. For one thing, the iron doesn't "like" reversing its magnetization, so there are some glitches in the process (hysteresis). For another, as you pour more and more power into the primary, the iron core starts to lose the ability to magnetize itself more and more to match. The result will be distortion, improper load on the valve, etc (saturation).

Many guitar amp gurus think that some of these flaws in cheap transformers actually sound good in a guitar amp.

I get a little bit of the sense that you are trying to run in three directions at once, i.e., there are multiple different hypothesis as to what is causing the problem, but they are not being systematically eliminated, one by one, until the answer is found.

I think the reason you're not getting too many replies from other people, is that there isn't quite enough information here for someone else to nail down the problem for you.

The starting point is for you to provide an exact description of the problem, and a complete schematic of the corresponding circuit. That will allow others to make plausible guesses as to the problem.

-Gnobuddy

 

[djgibson51]

"So is the amp is well behaved at all temperatures as long as its not set to full gain? If this is the case, you almost certainly do have instability, probably oscillation at frequencies too high to hear."

Humour me. Tack a bypass cap (300-1000pF) across one or both of the load resistors of the first tube and report back the result.

 

[Gnobuddy]

DJGibson is a wise man, and much thriftier with words than I am. Do try his suggestion!

-Gnobudddy