1st P/P build, questions

[shooter]

The top schematic is from a sunn 200s, the bottom an audiophile home-build.  The sunn uses a single bias while the other uses a separate bias on each tube.  Is there a practical reason for 2?  Is there a tonal difference?

Both schematics use NFB, on the sunn there is a 680R to ground on the signal side of the speaker.  Why?   I’ve seen schematics where the NFB can be switch in or out.  Would this be like a poor-mans “gain / boost” circuit?

Thanks for your help

dave

[ernie_jr]

the 680 ohm resistor is to provide some load if the amp is powered up without a speaker hooked up. it helps to protect the O/P and prevent flyback voltage spikes. As for the bias, either is fine. if you have 2 it does not matter as much if the tubes are not matched as each tube can be set on its own.
hope this helps,
ernie

[jjasilli]

1.  See:  http://www.mif.pg.gda.pl/homepages/frank/sheets/084/k/KT88_GEC.pdf  note that per this data sheet, ultralinear fixed bias operation requires 2 separate bias voltage sources withn a stated margin of error.  My Stromberg Carlson 2X KT-88 amps each have Bias Adjust & Bias Balance pots.  

2.  Why the 10Ω 1W cathode resistor?  My experience is that 10Ω 1/2W acts as a fuse.  1W seems to work against that purpose.  

[shooter]

Wow, thanks for the datasheet, the modern one I got sucks!  N i'm not sure about the 1w 10r, i'm leanin to the Sunn PA with no r in the cathode crkt. how Is the  r's wattage calc'd in class A/B?

I think i'll steal the 680r idea!, adjusting the watts as needed.

[jjasilli]

how Is the  r's wattage calc'd in class A/B?  Actually it isn't  10Ω 1/2 W is an old-fashioned standard -- as a safety fuse -- for the big bottle pentodes.  In case of an over-current condition, such as short circuit from a tube gone bad, the resistor blows. 

However, the wattage rating could be derived using Ohm's Law. 

Note that you can measure the tube's current draw with a voltmeter across the 10Ω.  These days this is usually done with a precision 1Ω 1W - 3W resistor (we don't want that resistor to blow).  But remember the 10Ω resistor will move the decimal point 1 place to the right, and give a reading 10X higher than the truth.  This can freak a person out!  So now if I use a 10Ω as fuse I also add a 1Ω is series, then measure across that.

Finally, if you use the 10Ω it's effectively adding some negative bias voltage, by making the cathode more positive relative to G1.  If you delete the 10Ω you'll need to add that much more fixed bias voltage.  The tube charts will give you your target (bogie) bias voltage for your set-up.  For first power-up:  Make sure you have the proper target bias voltage on the G1 socket lug with no tubes inserted.  Then try the circuit with power tubes installed, while reading current draw on a meter.  If there's insufficient bias, tubes may fry before you can react.  Hence it's a good idea to use 10Ω resistors or actual fuses at least temporarily for testing / first start-up.

[shooter]

I've attached my crude working schematic, I was wondering if I can steal the presence circuit from the attached fender schematic but i'm not sure where to "plug" it into my schematic.  does the presence takeout/addtoo the nfb or is it just a coincidence they both tie at the same place?  thanks

dave

[jjasilli]

Drawing a mental blank on your exact type of PI.  I'm pretty sure a list of PI schematics is posted somewhere on this Forum. 

Anyway your PI looks much like a "cathodyne" a/k/a "split load" PI.  Checkout the Fender Princeton schematic.  NFB is inserted into the cathode of the driver (first) stage of the PI.  The cathode resistance is split into two resistors, and NFB is inserted there. 

[shooter]

so, in theory will this section in red work?

I'm looking for a way to get a "presence" effect along with a variable nfb/gain  in the pre section.  I do have 1/2 a 12ax7 spare if there's a "better" cut-in for this basic pre.  thanks.

oh, the PI came from a '61 kt88 datasheet as a circuit supplement for a 30-100W basic amp.

[jjasilli]

Close but no cigar.    For 2 reasons: 

1.  A pot is probably ok. But a fixed resistor to ground should be under the pot.   That way you can't dial resistance to zero and kill the tube's bias. 

2. You didn't split the cathode resistor per the Princeton schematic.  NFB is AC signal. It will see the large cathode bypass cap in your drawing as a short circuit to ground.  All your NFB will bleed to ground and have no effect.  In the Princeton Reverb schematic, NFB is injected at the junction of the Bottom of the 1500R resistor + the top of the 47R resistor, which lifts NFB from ground.  The 47R resistor is part of a voltage divider, acting as the shunt resistor with the series resistor in the NFB loop.  What's left of NFB voltage passes up through the bypass cap and around the 1500R resistor.

If you want to use a pot, starting from the bottom up, maybe a 47R resistor to ground, then the 1500 resistor on top of that could be replaced with, say, an 820R resistor to preserve bias, then a 5K pot.

[shooter]

I get how the nfb hooks up in the Princeton, what I'm trying to get is how presence works, most of the schematics I've seen have it wired with nfb as an "input"? to the phase invertor.  since I'm trying to use a different than standard phase invertor, I'm looking for a "new place" to make that happen.  I've attach a schematic that I basically stole the presence circuit for starters, but I need a home to solder it to.  Thanks for the help.

[sluckey]

Presence is typically found in the NFB circuit. It usually consists of a pot and a capacitor used as a tone control.  A standard NFB circuit (without presence) will feed back all frequencies equally, reducing the gain of the power amp the same for all frequencies. Adding the presence pot/cap allows you to vary the amount of NFB such that all frequencies are not  fed back equally. For example, if you cut the high frequencies in the feedback signal, the power amp gain will not be reduced as much as the lower frequencies. The net result is the amp will sound brighter. So, a presence circuit behaves very much like an active tone control.

So, the idea of adding a presence control to a NFB circuit is not concerned with where the NFB is applied. Nor does it matter which phase inverter you use. It doesn't even matter if it is a single end amp. You should be able to get your standard NFB circuit working properly, then just add a pot/cap as a simple tone control in the loop. Probably a simple and fun experiment.

[jjasilli]

I get how the nfb hooks up in the Princeton, what I'm trying to get is how presence works, most of the schematics I've seen have it wired with nfb as an "input"? to the phase invertor.  since I'm trying to use a different than standard phase invertor, I'm looking for a "new place" to make that happen.  I've attach a schematic that I basically stole the presence circuit for starters, but I need a home to solder it to.  Thanks for the help.

Sorry maybe my post wasn't clear.  See the last sentence: "If you want to use a pot, starting from the bottom up, maybe a 47R resistor to ground, then the 1500 resistor on top of that could be replaced with, say, an 820R resistor to preserve bias, then a 5K pot."  Schematic attached.

[tubeswell]

so, in theory will this section in red work?

Yes your presence control will work to the point that as you short across the pot, you will remove the NFB as well. Seeing as how Rk is unbypassed in your schematic, you will still have an element of cathode current feedback providing a cleaner signal and less gain overall in the stage, than what you otherwise would get if the Rk was bypassed. And the pot will have DC on it from tube current for the driver stage that the NFB is feeding into, so it will be scratchy when you turn it

[jjasilli]

I continue to disagree.  In the red en-circled drawing, all NFB will bleed to ground through the cap; the pot is rendered superfluous to control NFB.  OTOH, the pot will change the operating point of that tube stage, which may indirectly affect tone. 

Another useful NFB circuit for this purpose is the Fender 5E7 Bandmaster, which has both NFB and a pretty similar PI:     http://www.el34world.com/charts/Schematics/files/fender/bandmaster_5e7_schem.pdf

Note that the Bandmaster's Presence control is a type of simple, treble bleed tone control within the NFB loop.  To the extent that more NFB treble is bled to ground, it fails to attenuate the main signal.  This allows NFB to affect speaker cone excursion in the lower frequencies, where it counts, while making treble adjustable. 

Also, note that the Bandmaster's NFB gain stage has no bypass cap.  If a bypass cap were simply added, once again all NFB would bleed through it to ground.  The solution to that issue is in the Princeton circuit.

[ernie_jr]

SUNN also is UL and uses a cap and resistor in parallel.
Ernie

[terminalgs]

I continue to disagree.  In the red en-circled drawing, all NFB will bleed to ground through the cap; the pot is rendered superfluous to control NFB.

depends 100% on the value of the cap. a 25uf cap, yes. for a .01 cap, no.

with a smaller cap,  like .01,,  high frequencies will be lost to ground, low frequencies will survive.

Those low frequencies will make it to the cathode and then to the plate.  If shooter's  triode is situated in the circuit such that the surviving low frequencies are out-of-phase with the input signal, when they meet at the plate, that frequency will  be attenuated by some small percentage.  IF they are in-phase,,  the amp will howl as it'll actually be positive feed-back, like a mic in front on a PA speaker.

OTOH, the pot will change the operating point of that tube stage, which may indirectly affect tone. 

yup, and that's the chief reason not to do it.
 
It might work if the nfb comes in on the center lug of the pot.  depends on resistor,pot,cap values, and the rest of the circuit.

ditto what sluckey said.   

[shooter]

Thanks guys, I see my errors!   The bandmaster schematic is what I've been hunting for.  It's in parallel with, not series with the Rk! Thanks SLUCKEY for confirming what I thought presence was electrically, a band pass/trap filter.  Just fired up my SE quad el84, it does in fact amplify!  Now I gotta get the numbers, scope the signal in and see how well it amplifies! Then fix a valvstate 2x65, have holidays and finally get parts for this P-P!  I love winter in the north country!

again, thanks

dave

[jjasilli]

terminalgs:  depends 100% on the value of the cap. a 25uf cap, yes. for a .01 cap, no. . . with a smaller cap,  like .01,,  high frequencies will be lost to ground, low frequencies will survive.

Yes, BUT: i)  a cap that small will defeat the gain boost of a "standard" value bypass cap; and ii) the cap in parallel with a resistor (or variable resistor) creates a Notch Filter; but shooter wants a Presence control which affects treble.

shooter:  Thanks guys, I see my errors!  The bandmaster schematic is what I've been hunting for. OK, but remember, in your stock schematic the 1st (driver) stage of your PI has a cathode bypass cap which significantly boosts the gain of that stage.  To preserve that, AND inject NFB, WITH a Presence control, you need something like the Bandmaster Presence circuit, injected into something like the Princeton split cathode resistor circuit.

[terminalgs]

terminalgs:  depends 100% on the value of the cap. a 25uf cap, yes. for a .01 cap, no. . . with a smaller cap,  like .01,,  high frequencies will be lost to ground, low frequencies will survive.

Yes, BUT: i)  a cap that small will defeat the gain boost of a "standard" value bypass cap; and ii) the cap in parallel with a resistor (or variable resistor) creates a Notch Filter; but shooter wants a Presence control which affects treble.

I only meant to say that I didn't think your statement "all NFB will bleed to ground through the cap" was really accurate.  some NFB frequencies would bleed to ground, some would not,,  those would be applied as NFB,,  in a presence tone control manner (albeit flawed by issues shooter has already corrected). 

pretend the 5K pot isn't a pot, its a 5K resistor.  how different is it compared with the presence control in a bassman 5F6A or a Bandmaster 6G7 (other than the LTPI vs. cathodyne diff),   (1) the 5K resistor is the R1 in the R1/(R1+R2) NFB voltage divider and (2) the parallel cap is a frequency dependent tone shaping cap. 

(maybe I'm wrong, if so, I wanna know for sure!  apologies, I'm a stubborn taurus)

[tubeswell]

Yes the pot bypass cap will shunt some of the AC in the NFB to ground, depending on the size of the cap. The pot (wired as a variable resistor) in parallel with this cap will control the how much of the 'remaining' (i.e.; 'unshunted') AC is present at the knee of the voltage divider in the NFB loop (note that we can't see the other NFB resistor in the schematic). As the resistance across the pot is progressively diminished with pot rotation, the amount of NFB getting to the cathode of the affected stage will be reduced - until at zero rotation where no NFB will be present.

In addition, the same gain stage will also be affected by the cathode current feedback originating from the unbypassed cathode resistor (Rk) (which is in series with the knee of the voltage divider in the NFB loop and the stage's cathode). The cathode current feedback from the unbypassed Rk will already be 'cleaning up' the signal somewhat. And furthermore, because there is no cap between the pot and the Rk for the gain stage, there will be DC on the pot resulting from the tube current for the gain stage, which will cause scratchiness on the pot when it is rotated.

[jjasilli]

terminalgs: how different is it compared with the presence control in a:

bassman 5F6A: in the bassman, the treble bleed cap is from Wiper (lug 2) to Ground, not across the whole pot (from lug 1 -3) as in shooter's red encircled drawing.  I'm not sure what actual difference this makes.  It may only affect the functional taper of the pot.  In both cases with the pot at full resistance, or zero resistance, the effect should be the same for each circuit.  In each case, at zero resistance all NFB will bleed to ground; so NFB can be turned off.

Bandmaster 6G7:  here the NFB treble bleed is standard, with a pot sitting atop a cap to ground (though somewhat complicated because this treble bleed R-C circuit is in parallel with the 1500Ω resistor -- I suspect this makes little difference to the treble bleed effect.)  Here NFB cannot be fully turned off.  With the pot at zero resistance, the cap will still lift some NFB from ground.