Page 21 of 54

#301 Re: Remember These?

Posted: Fri Jan 04, 2019 8:50 pm
by Ray P
Cressy Snr wrote: Fri Jan 04, 2019 2:38 pm Audio act schematic with pin numbers for 13E1:
Thanks Steve. I always have this gnawing doubt creep into my mind when it comes to this sort of thing; dunno why but I do!

#302 Re: Remember These?

Posted: Fri Jan 04, 2019 8:51 pm
by IslandPink
It wasn't obvious to me at first, so you're not different ! :D
I could try and explain it so : if you have the feedback connected, it's not like the higher gain is there, then the feedback reduces it - there' no gap between the two things - the feedback is already applied in nanoseconds, so the higher swing is not 'seen' at the 13E1 grid.
Also... the gain of the 13E1 from grid to cathode is not any different with & without the feedback. Hence you only need the grid bias that is consistent with the output power from the 13E1.

#303 Re: Remember These?

Posted: Fri Jan 04, 2019 10:17 pm
by Cressy Snr
My poor brain can't compute this.

If I make a single ended amplifier that has a gain of 100 from end to end then apply feedback. I have to shove a lot more voltage in at the input to get the same volume I would have got without feedback. Gain has gone down, so it must have gone somewhere.

If in a negative feedack system you have to shove more up the input, compared to the open loop situation, then whether the output valve sees that signal or not, you still have to provide it don't you? I provided the high drive; are we saying that, that drive provision is unnecessary, I think it is very necessary, otherwise it just makes a nonsense of everything I learned. With negative feedback, you have to put more in to get the same out. Why is there a feedback equation if not? :?

#304 Re: Remember These?

Posted: Fri Jan 04, 2019 10:41 pm
by Nick
Yes, the key as Mark said is open loop gain and the feedback factor. Most of it is here

Image

As Mark said, the output valve doesn’t know about feedback, all it knows are its grid and cathode, so the voltage swing on the grid is the voltage swing that needs to produce the required output into the load. If you ran it open with no feedback and supplied the grid with the same signal you would get the same output as if you had feedback. The feedback allows the system to apply the correct signal to the grid of the output valve such that its cathode produces the required result.

In Steves example of the single ended amp with a open loop gain of 100 (40dB), then if you apply 30dB of feedback, the overall gain of the system will be 10dB, the 30dB fain of the feedback will consume 30dB of the system gain.

This is where it gets problematic with parasitics and group delay. If HF is not part of the feedback because of losses in the loop then at HF the system as the original gain of 40dB at HF as there is nothing to consume the gain. If there is a delay in the system (as there will be), then at some frequency the feedback will change from being negative, to being less negative, and then become positive.

Nothing is ever as simple as we might hope.

But back to this amp, the output valve load lime is a 8 ohm cathode follower loadline, and that will show you what signal on the grid will be needed to produce the signal on the cathode. Of course headroom is good, but how much is the important factor. What rearly interests me (and is why I would like to attach a scope to the thing) is will the grid be driven positive into A2 and with the grid circuit you have what happens then...

#305 Re: Remember These?

Posted: Fri Jan 04, 2019 10:43 pm
by Nick
you have to put more in to get the same out. Why is there a feedback equation if not?
Yes, but the summing point is not the grid of the output valve, thats inside the loop.

#306 Re: Remember These?

Posted: Fri Jan 04, 2019 10:54 pm
by Cressy Snr
Nick wrote: Fri Jan 04, 2019 10:43 pm
you have to put more in to get the same out. Why is there a feedback equation if not?
Yes, but the summing point is not the grid of the output valve, thats inside the loop.
:idea1: Gotcha now :)
There's no need for the driver to produce wall to wall swing, only what's needed by the output stage. Every day's a schoolday :oops:

I now get why Bruce uses a simple cascade with a 6N something P twin triode, on the Mini-Beast. Both sections are identically configured.
I've just clip leaded out one of the 220R resistors on the driver, so its now identical to the input stage. Works perfectly well.

#307 Re: Remember These?

Posted: Fri Jan 04, 2019 11:09 pm
by IslandPink
I'll try & translate :D
You feed a signal into the input valve grid ; as you do so, the cathode (due to feedback) is going up and down with it - quite a bit - as the feedback is applied. So, you don't get the gain you expect from the first valve.
After that, gain-wise, everything operates as expected if there wasn't feedback.

Sorry, I missed your extra comments - but you are obviously catching up now. Good job !

#308 Re: Remember These?

Posted: Fri Jan 04, 2019 11:13 pm
by Cressy Snr
IslandPink wrote: Fri Jan 04, 2019 11:09 pm I'll try & translate :D
You feed a signal into the input valve grid ; as you do so, the cathode (due to feedback) is going up and down with it - quite a bit - as the feedback is applied. So, you don't get the gain you expect from the first valve.
After that, gain-wise, everything operates as expected if there wasn't feedback.

Sorry, I missed your extra comments - but you are obviously catching up now. Good job !
You know what this means now: another blasted schematic. :lol:
Thank God it's on the computer this time.

#309 Re: Remember These?

Posted: Fri Jan 04, 2019 11:16 pm
by IslandPink
Yowsa !
We love your schematics !

#310 Re: Remember These?

Posted: Fri Jan 04, 2019 11:18 pm
by Cressy Snr
I would suspect you don't want to be trying to drive the grid of the 13E1 positive, with cap coupling from the driver stage :shock: Another argument for sufficient drive and no more.

#311 Re: Remember These?

Posted: Fri Jan 04, 2019 11:32 pm
by Cressy Snr
Nick wrote: Fri Jan 04, 2019 10:41 pm What rearly interests me (and is why I would like to attach a scope to the thing) is will the grid be driven positive into A2 and with the grid circuit you have what happens then...
I could bring it to your place sometime this month, before I put it in a box.

#312 Re: Remember These?

Posted: Sat Jan 05, 2019 8:20 am
by Ray P
Cressy Snr wrote: Fri Jan 04, 2019 11:32 pm
Nick wrote: Fri Jan 04, 2019 10:41 pm What rearly interests me (and is why I would like to attach a scope to the thing) is will the grid be driven positive into A2 and with the grid circuit you have what happens then...
I could bring it to your place sometime this month, before I put it in a box.
It'll be interesting to see what Nick makes of it...

#313 Re: Remember These?

Posted: Sat Jan 05, 2019 10:12 am
by Nick
Cressy Snr wrote: Fri Jan 04, 2019 11:32 pm
Nick wrote: Fri Jan 04, 2019 10:41 pm What rearly interests me (and is why I would like to attach a scope to the thing) is will the grid be driven positive into A2 and with the grid circuit you have what happens then...
I could bring it to your place sometime this month, before I put it in a box.
That would be good.

--
Nick

#314 Re: Remember These?

Posted: Sat Jan 05, 2019 10:14 am
by Nick
Cressy Snr wrote: Fri Jan 04, 2019 11:18 pm I would suspect you don't want to be trying to drive the grid of the 13E1 positive, with cap coupling from the driver stage :shock: Another argument for sufficient drive and no more.
That was what I was thinking. The CCS will help keep it under control to some extent.

#315 Re: Remember These?

Posted: Sat Jan 05, 2019 2:13 pm
by Cressy Snr
OK, here are some boring graphs of the amp driving an 8 Ohm dummy load.

There's barrel distortion on my iPad camera, but it should give the general idea,
Image

1KHz sine wave