Cheapest power supply transformers?

[jim1976]

I think a new thread might be a good idea, I just started one about output transformer measurements and you really seem to have the gear for looking at this sort of thing!

So did you run the spectrum analyser at any earlier stages in the same amp? its just the 2nd and 3rd seem high for such low outputs maybe an earlier stage is clipping before the power tubes? did you have a scope on it at the same time to see if there was clipping? was it a "good sounding 5e3"?

Thats a great bit of gear you have there btw

Thanks

[jim1976]

But is the distortion from the gain stages, over driven inputs? is it amplified clipping?

[Steve Conner]

All tubes have non-linear transfer characteristics. The graph of output voltage vs. input is a curve, not a straight line. It is a mathematical fact that that generates harmonics. A single 12AX7 stage will produce a few percent of harmonic distortion, mostly 2nd, when driven to a healthy output level. The McIntosh uses a huge amount of negative feedback to minimise it.

Clipping is just an extreme form of non-linearity. The transfer curve has a sharp knee in it that generates lots of high-order harmonics.

[tubeswell]

Your a kiwi cool! have you had a play with any Jansens by chance?

I've owned several of their amps and guitars over the years.

Did you read that chapter in the link I posted?

[Tom Phillips]

Hi James,
I only have time for short answers right now. Here they are

...So did you run the spectrum analyser at any earlier stages in the same amp? ...

Not during these 5E3 tests

...its just the 2nd and 3rd seem high for such low outputs maybe an earlier stage is clipping before the power tubes? ...

Remember that a logarithmic vertical scale is shown. The second harmonic is not half the amplitude of the fundamental. It is less than one thousandth! (I upgraded the attachment in post #74 with additional info)

...did you have a scope on it at the same time to see if there was clipping?...

Of course and no stage of the amp was clipping. I have companion time domain photos too but no time to find, prep and post right now.

... was it a "good sounding 5e3"?...

Yes

...Thats a great bit of gear you have there btw...

Yep. I’m a test equipment freak. It’s nice to be able to do investigations like this but it does eat up lots of time compared to a straight on repair. Especially when the results are documented well enough to be sorted out and understood at a later date. Being able to see the signal in the frequency domain is somewhat like a look into another dimension.

Cheers,
Tom

[R.G.]

One thing I do know about is sensory evaluation, I'm a winemaker and I can tell you it's not accurate its affected by so much mood to how hungry you are.

I am thoroughly impressed with you for making with that statement. It's absolutely correct, but apparently the majority of humanity cannot accept that concept as being true. Perhaps especially about music, but taste and smell are about the same.

What our sensory nerves tell us are heavily, heavily filtered by several layers of completely unconscious processing before our conscious minds get a chance at the information.

[Chuck H]

Well I'm having NO trouble getting on that train. I go to great pains to make my personal amps exactly how I want them. And on some days I love them. I can actually play guitar without the distraction of evaluating the tone. Which is, to my mind, a great and rare accomplishment as a designer. But... Other days I just can't get a good sound out the same amp!?! This happens with all things created equal as the hardware goes. I've tested operating parameters and wall AC, etc. The only possible difference between when I like my tone and when I don't is my state of mind.

[tubeswell]

It isn't, watch this HP 3580A Spectrum Analyzer and Harmonic Distortion on the McIntosh MC275.MOV - YouTube
Plate resistor load and bias point only move it down the scale, do you have test gear? I really think its the only way to tell.

I'm a hobbyist who's not rich enough to own a spectrum analyser, but you don't need to own one to understand how tube harmonic distortion works. Merlin Blencowe explains it it that chapter in the link I posted. (Note that there are a couple of typos in that edition of the chapter). A while back I did these plate characteristics charts for 12AU7 triode with different plate load and bias points.



You'll see that in both examples there is unclipped harmonic distortion present. See how the grid curves are not evenly spaced? As the grid voltage change gets more negative, the amount of plate current change is proportionately less than if the grid voltage change is happening closer to the Vg=0 end of the load line. You can see this is more exaggerated in the example running the 20k plate resistor and the 1k8 cathode resistor - the stage is biased to idle at Va = 232V, with Vg = -10V. A change of -10V on the grid (i.e. to Vg= -20V) results in a plate voltage change of 89V (i.e. to 321V), whereas a change of +10V on the grid (i.e. to Vg= 0V) results in a plate voltage change of 136V (i.e. to 96V). So the voltage swing at the plate is distorted compared to the voltage swing at the grid. That's where the harmonic distortion comes from. Do you understand now?

[jazbo8]

I'm a hobbyist who's not rich enough to own a spectrum analyser, but you don't need to own one to understand how tube harmonic distortion works. Merlin Blencowe explains it it that chapter in the link I posted. (Note that there are a couple of typos in that edition of the chapter). A while back I did these plate characteristics charts for 12AU7 triode with different plate load and bias points.

[ATTACH=CONFIG]21429[/ATTACH]

You'll see that in both examples there is unclipped harmonic distortion present. See how the grid curves are not evenly spaced? As the grid voltage change gets more negative, the amount of plate current change is proportionately less than if the grid voltage change is happening closer to the Vg=0 end of the load line. You can see this is more exaggerated in the example running the 20k plate resistor and the 1k8 cathode resistor - the stage is biased to idle at Va = 232V, with Vg = -10V. A change of -10V on the grid (i.e. to Vg= -20V) results in a plate voltage change of 89V (i.e. to 321V), whereas a change of +10V on the grid (i.e. to Vg= 0V) results in a plate voltage change of 136V (i.e. to 96V). So the voltage swing at the plate is distorted compared to the voltage swing at the grid. That's where the harmonic distortion comes from. Do you understand now?

Jim,

Please read this: The Distortion Factor, or the headline says it all - "The monolithic view of distortion, namely clipping, is false. There are many more forms of distortion than merely clipping, although clipping is quite important, if not the most important."

Jaz

[Jazz P Bass]

In it's simplest form, when comparing the input to the output, ANY change in the waveform is regarded as distortion.

[jim1976]

All tubes have non-linear transfer characteristics. The graph of output voltage vs. input is a curve, not a straight line. It is a mathematical fact that that generates harmonics. A single 12AX7 stage will produce a few percent of harmonic distortion, mostly 2nd, when driven to a healthy output level. The McIntosh uses a huge amount of negative feedback to minimise it.

Clipping is just an extreme form of non-linearity. The transfer curve has a sharp knee in it that generates lots of high-order harmonics.

So if you have a spectrum analyser and an amp with a negative feed back would it be possible to show us a shot with and with out the negative feed back? and perhaps a shot before the phase inverter to show the progression?

[jim1976]

I'm a hobbyist who's not rich enough to own a spectrum analyser, but you don't need to own one to understand how tube harmonic distortion works. Merlin Blencowe explains it it that chapter in the link I posted. (Note that there are a couple of typos in that edition of the chapter). A while back I did these plate characteristics charts for 12AU7 triode with different plate load and bias points.

[ATTACH=CONFIG]21429[/ATTACH]

You'll see that in both examples there is unclipped harmonic distortion present. See how the grid curves are not evenly spaced? As the grid voltage change gets more negative, the amount of plate current change is proportionately less than if the grid voltage change is happening closer to the Vg=0 end of the load line. You can see this is more exaggerated in the example running the 20k plate resistor and the 1k8 cathode resistor - the stage is biased to idle at Va = 232V, with Vg = -10V. A change of -10V on the grid (i.e. to Vg= -20V) results in a plate voltage change of 89V (i.e. to 321V), whereas a change of +10V on the grid (i.e. to Vg= 0V) results in a plate voltage change of 136V (i.e. to 96V). So the voltage swing at the plate is distorted compared to the voltage swing at the grid. That's where the harmonic distortion comes from. Do you understand now?

Thanks for that Tubeswell theres a lot to get through there!

[jim1976]

But what type and how much of what is good and where should the distortion take place, spectrum analysis is the only way to find out.

[jim1976]

But what type and how much of what is good and where should the distortion take place, spectrum analysis is the only way to find out.

[jim1976]

In it's simplest form, when comparing the input to the output, ANY change in the waveform is regarded as distortion.

That is spot on! and everything amplified is distorted.