possible advantage to global??? - diyAudio
Go Back   Home > Forums > Amplifiers > Tubes / Valves

Tubes / Valves All about our sweet vacuum tubes :) Threads about Musical Instrument Amps of all kinds should be in the Instruments & Amps forum

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 25th December 2010, 05:59 AM   #1
diyAudio Member
 
Join Date: Jun 2004
Location: Indiana
Default possible advantage to global???

I have to switch over to graveyard shift for one night tomorrow so I am pulling some late hours at the computer to switch over my body clock. I was running some spice sim's for my edification and noticed something that I think may point to an advantage of gNFB over local.

We are all familiar with some of the advantages of local FB on the output stage but as I was playing I noticed that the local FB (plate to plate) by virtue (or vice perhaps) of reducing the gain of the output stage reduced the maximum possible output power considerably because driving the grid right up to 0V resulted in less output voltage swing given the same OPT primary impedance.

With gNFB the gain of the output stage is not reduced and so full output is still available as long as the input signal can be increased enough to compensate for the loss in overall gain.

This brought a couple of questions to mind.

1) Shouldn't this have been obvious to me without having to run simulations?

2) Is it therefor customary to use a different OPT impedance for local feedback applications?
__________________
If We The People refuse to hear the truth we will be ruled by liars.
  Reply With Quote
Old 25th December 2010, 07:33 AM   #2
diyAudio Member
 
artosalo's Avatar
 
Join Date: Feb 2010
Quote:
reducing the gain of the output stage reduced the maximum possible output power considerably because driving the grid right up to 0V resulted in less output voltage swing given the same OPT primary impedance.
How much local NFB you had ?
How much max. output was decreased due to local NFB ?

What does your similation give if you lower the g1 voltage ?
  Reply With Quote
Old 25th December 2010, 08:37 AM   #3
diyAudio Member
 
Join Date: Jun 2004
Location: Indiana
IIRC it was about 14dB. I did not try varying the screen voltage.
__________________
If We The People refuse to hear the truth we will be ruled by liars.
  Reply With Quote
Old 25th December 2010, 06:47 PM   #4
diyAudio Member
 
Sch3mat1c's Avatar
 
Join Date: Jan 2003
Location: Milwaukee, WI
Send a message via ICQ to Sch3mat1c Send a message via AIM to Sch3mat1c
Maximum power output is at clipping, not an arbitrary voltage threshold. A threshold which depends on how the circuit is wired. Vgk = 0 isn't a magical point of clipping; that depends on your load resistance (which you usually pick so it is). Class 2 amplifiers, obviously enough, don't clip at Vgk = 0; most class 1 amplifiers do, at or below 0V.

Depending on what resistor values you used, your driver might not even be able to supply enough voltage to drive the output to clipping.

FYI, local feedback is preferred because it improves bandwidth and phase margin independent of the number of stages. With the improved response per stage, you could then cascade stages and add more NFB around groups of stages or globally, with much improved performance over the case of the same total amount as gNFB.

Tim
__________________
Seven Transistor Labs
Projects and Resources
  Reply With Quote
Old 25th December 2010, 08:47 PM   #5
diyAudio Member
 
tubelab.com's Avatar
 
Join Date: Jul 2005
Location: West Virginia panhandle
My experience with the red board reveals no major difference in maximum power output with no feedback, GNFB, or local Schade feedback. Given a driver capable of feeding the output tube all it needs I find only 3 things that limit the maximum power output. The choice of output tube. Obviously plate dissipation and voltage ratings are important, but so is peak current carying capability and saturation characteristics. For a given tube the only variables are load impedance and plate voltage.

The use of local feedback will indeed reduce the gain of the output stage. The driver tube must be capable of supplying the increased drive voltage demands without undue distortion. The use of GNFB will reduce the gain of the entire amplifier. In either case 14 db of feedback will reduce the gain of the entire amp by 14db. With GNFB the gain of the driver and the output stage is reduced. The amount of reduction in each stage depends on the individual stage gains. With Schade feedback all 14db of gain reduction will come from the output stage, demanding 14db of additional headroom in the driver.

14db is probably too much to ask from the driver stage. My experiments with the red board use 6 to 8 db of Schade, GNFB, both, or neither. I find that the amp will measure about the same using local or GNFB. I get better distortion numbers from idle to just below clipping with both feedback paths enabled. In listening tests with dynamic music I find Schade feedback only preferrable. GNFB alone is OK, but both feedbacks together take all of the life out of the music.
__________________
Too much power is almost enough! Turn it up till it explodes - then back up just a little.
  Reply With Quote
Old 25th December 2010, 09:26 PM   #6
diyAudio Member
 
artosalo's Avatar
 
Join Date: Feb 2010
Quote:
With GNFB the gain of the driver and the output stage is reduced.
I think this is not the case in typical GNFB arrangement.
Can you give an example where the above takes place?

I have documented one of my PP-amplifiers with 14 dB GNFB and without.
The amplifier is conventionel with 6L6 PP-output stage, 6N7 long tailed pair phase splitter and 6SL7 voltage amplifying stage.
GNFB is fed normally from the secondary of the OPT to the cathode of the voltage amplifying stage.

The AC-voltage measurements show that when GNFB is activated, the gain of the voltage amplifying stage only is reduced from 32 dB to 18 dB, i.e. by the amount of GNFB.
The gain of all the other stages remain unchanged.
  Reply With Quote
Old 25th December 2010, 10:55 PM   #7
diyAudio Member
 
Join Date: Jun 2004
Location: Indiana
Something else is going on with my simulation that is making it hard to determine exactly what is going on.

I realized this morning that I was accidentally running with the outputs connected triode mode but I was getting 15 or so watts (5V+ peak into 4 ohms) before clipping which occured when the driver output voltage was equal to the bias voltage of the outputs. Clearly this is way too much power from trioded EL84s.

This morning I redrew it for pentode mode. This allowed over 10V peak into 4 ohms. Again way to much. Something is goofy either with the Koren models I am using or my transformer model or something.

The topology is Floating paraphase with 12AT7s using 68K plate load and about 1.6V cathode bias (680 ohms unbypassed). The output stage has 270 ohm resistor bypassed by 220uf caps on each EL84 cathode (10.5V cathode voltage). Power supply is 300V with the screen taken from the same B+ with 1K dropping resistor. The DC voltage drop at the screen is only about 1V.

The OPT is simulated with the two sides of the primary spec'ed at 60H and 1ohm DCR. Secondary is 0.125H and 0.1ohm DCR. K=1. My understanding is that this should be a bit over 8K P-P anode load.

I ran the simulation to get the maximum unclipped output then I added 270K plate to plate to my original model (this was triode connected remember) the output level for the same input dropped by a factor of five but the signal to the EL84 grids was still at the maximum before clipping (about 10V peak) so any further increase in input voltage just drove the outputs into clipping with no increase in output.

I switched to gNFB by dividing the input tube's cathode load into two with an upper R of 570 ohms and a lower of 130 ohms (using 110 ohms unbalanced the bias due to the parallel Rfb). I adjusted the value of Rfb to get the same overall gain reduction and ended up at, IIRC, about 760 ohms for Rfb. In this case the gain of the input/PI stage was reduced so that the input to the EL84 grids was reduced by a factor of 5. For this reason I could increase the input voltage to the point were the grids of the EL84 were again driven to the maximum 10V. The output power was now restored to its original value.

So I have two things going on. One something is wrong with the power output levels that I am getting and two is that the local feedback is reducing the total available power out. I suppose that I could use a follower between the PI and the outputs to allow some grid drive but this would also increase the potential output without FB.

BTW, I am calculating the output power by (Vpeak*1.4)^2/4

Any enlightenment you can offer would be appreciated.
__________________
If We The People refuse to hear the truth we will be ruled by liars.

Last edited by mashaffer; 25th December 2010 at 10:58 PM.
  Reply With Quote
Old 25th December 2010, 11:20 PM   #8
Jeb-D. is offline Jeb-D.  United States
diyAudio Member
 
Jeb-D.'s Avatar
 
Join Date: Nov 2005
Location: SoCal
It's been my experience that local feedback (plate-to-plate) actually decreases the driver stage gain, not the output stage gain. The output stage still requires the same amount of drive voltage and it's power output should remain appox equal to open loop.

The potential problem which sounds like you are experiencing is that as you increase the feedback ratio, the AC impedance seen by the driver proportionally decreases (wich is why gain decreases). There comes a point where the feedback ratio is so high, the driver is unable to swing the required voltage (because its working into an impedance that is too low).

Last edited by Jeb-D.; 25th December 2010 at 11:28 PM.
  Reply With Quote
Old 25th December 2010, 11:47 PM   #9
Account disabled at member's request
 
Join Date: Apr 2009
I have no experience with simulation whatsoever, however the specs of the OPT seem strange to me. Normally it is specified in primary / secondary impedance or winding ratio.
60H per primary side to 0.125H for the secondary is 7k68 plate - plate primary impedance to 8 ohms secondary.
However 1 ohm DCR per primary side is far beyond reality; that should be at least 60 to 70 ohms for a good quality OPT, but maybe this has nothing to do with the sim?
  Reply With Quote
Old 26th December 2010, 09:46 AM   #10
diyAudio Member
 
artosalo's Avatar
 
Join Date: Feb 2010
Quote:
Any enlightenment you can offer would be appreciated.
If you had a real amplifier with such phenomenon, it would be possible to give some opinion about the reasons for unproper operation.

But I see that is quite useless to try to guess what is wrong with your simulation since there is some one thousand possibilities.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
What's the advantage of E-linear Michael Koster Tubes / Valves 34 30th August 2009 02:38 AM
Another advantage of IB mashaffer Subwoofers 2 15th September 2007 05:57 PM
What advantage from short arrays? Dumbass Multi-Way 15 17th November 2005 11:04 PM
advantage preamp seahag Chip Amps 9 4th October 2005 02:23 AM
Any advantage to Current Amplifiers? Rarkov Solid State 8 14th November 2002 09:59 PM


New To Site? Need Help?

All times are GMT. The time now is 09:48 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2