Gibson GA80 Vari Tone - Which PT and OT ?

[kagliostro]

Few time ago I posted a schematic I drawn of a Gibson GA80 (Vari Tone) Lite



Now I would like to investigate a bit more on this circuit

the power tubes that are used are 6L6GB (I don't know very much this tubes on details)

the original schematic is this



if I read correctly the voltage chart, the plates of the power tubes are supplied with 438v (B+) - 420v measured at plates and 360v on G2 (is this right ??)

the data about this amp say it is a 25W amp, so, with this voltage which will be the rigt OT primary impedance ?

looking to the datasheet I'm not able to establish a correct a-a impedance value for this voltages

http://www.mif.pg.gda.pl/homepages/frank/sheets/049/6/6L6GB.pdf

Thanks for any answer

K

[shooter]

Here's another data sheet, that doesn't help much either since the tubes are running over the 400vdc max plate, but I would think you would be ok at 6.6K   

It appears the more volts the more impedance so 8k may not be out of the question

http://tubedata.milbert.com/sheets/127/6/6L6GB.pdf

[DummyLoad]

use fender super reverb/bassman PT & OT.


--pete

[kagliostro]

Thanks Shooter & Pete

Pete, are this the transformers you mean ?

http://www.classictone.net/40-18029.pdf

http://www.classictone.net/40-18001.pdf

but ..... 4.2k as primary impedance of the OT 

on the datasheet seems that increasing voltages the primary impedance is to be increased 

http://www.mif.pg.gda.pl/homepages/frank/sheets/093/6/6L6GC.pdf


to me is also odd that the amp is declared as to be a 25W amp, but on the datasheet with 400v B+ in AB1 you get 55W 


 and looking closely to G2 PS also the 25k resistor connected to the cathodes is odd 

Franco

[DummyLoad]

works for fender, it'll work for you in this application.


if you want less power, use 6K6 or 5K with lower B+


IMO the bassman/super reverb is the path to follow. tremolux/vibrolux iron will work as well...

either of these:
http://www.tube-town.net/ttstore/Transformer/Outputtransformer/Fender/Hammond-1750J-Fender-Outputtransformer-125A6A-022848-Tremolux::2159.html
http://www.tube-town.net/ttstore/Transformer/Outputtransformer/Fender/Hammond-1760J-Fender-Outputtransformer-125A6A-022848-UPGRADE-Vibrolux::3974.html

with this PT:
http://www.tube-town.net/ttstore/Transformer/Power-Transformer/Fender/Hammond-290CEX-Fender-Power-125P26A-Vibrolux::2150.html

---OR---

either of these:
http://www.tube-town.net/ttstore/Transformer/Outputtransformer/Fender/Hammond-1760K-Fender-Outputtransformer-125A9A-022855-UPGRADE-Super-Reverb::5720.html
http://www.tube-town.net/ttstore/Transformer/Outputtransformer/Fender/Hammond-1750K-Fender-Outputtransformer-125A9A-022855-Super-Reverb::2182.html

with this PT:
http://www.tube-town.net/ttstore/Transformer/Power-Transformer/Fender/Hammond-290DEX-Fender-Power-125P5D-Super-Reverb::2151.html

--pete

[kagliostro]

Thanks Pete

Franco

[PRR]

> the plates of the power tubes are supplied with 438v (B+) - 420v measured at plates and 360v on G2 (is this right ??)

438V out of the rectifier. 420 at 6L6 Plates (18V lost in OT), but 30V at 6L6 Cathodes. There is 390V across the tube, 330V G2 to Cathode.

30V across 250r means 120mA cathode current. Self-bias, so we can not go to a low load impedance: plate current would rise but that would increase the bias and limit the plate current.

For 6L6 types, most of that 120mA goes to plate. 390V*0.120A= 46.8 Watts in two plates, 23 Watts per plate, a hair high of the 6L6GB's 22W rated limit. If we assume 5% G2 current then it is 22.23 Watts per plate. For genuine old 6L6GB, this is hot. Any modern "6L6" will be on the GC chassis rated 30W Pdiss.

There is not an exact match in the classic 6L6 datasheets, but P-P AB1 at 360V and 6,600 load is very close, and promises 26.5 Watts at 2% THD.

As a rough-guess for self-biased push-pull: 390V/0.120A is 3250 Ohms, P-P load should be twice this, 6,500 Ohms.

I have built an any-tube amp on a modified Ampeg with 400V supply 4K G2 resistor and 250r cathode resistor, 6.6K loading. It made 18W-23W (depending on line voltage) pretty clean. With 438V to start with it would be 25 Watts.

If you changed to self-bias and stiffened-up the G2 supply, then you *could* go down to 4K load, get more power, and (when maxxed-out) a very different amp.

[Willabe]

Is the 25K R from the screens to the K's correct?


               Brad   

[tubenit]

There should be a 100k between the .01 after the 5879 and the .002 into V1-7.

With respect, Tubenit

[kagliostro]

Ciao Tubenit

I've thinked about that before to draw my schematic and I ended that the 100k resistor was there as a mixer resistor

see the other 100k resistor present on the other channel between the .0047 and the .002 cap

so I decided that the presence of a residual 100k resistor between the .01 and the .002 cap was unuseful

what is to be tested is the value of the two capacitors  that now results to be in series (the .01 and the .002), the absence of the 100k

resistor changes the spec of the circuit, putting the two cap in strictly contac


---


About the 25k that is connected between G2 and ground (via the cathode resistor)

I think to it as a shunt resistor that is there as part of a resistor voltage divider formed by the 4k resistor and the 25k resistor

whose resulting voltage is leveled using the 10uF cap

What is misterious to me is why they connected the 25k resistor to ground via the cathode resistor instead of directly to ground

May be the fluctuation of voltage on the cathode are used as to perform a fluctuation on G2 voltage in according with the cathode

exploiting the varying ground reference voltage 

Franco

[shooter]

May be the fluctuation of voltage on the cathode are used as to perform a fluctuation on G2 voltage in according with the cathode

That's the way I'm seeing it also, I think the cathode is out of phase with G2 so it's acting like NFB? 

[sluckey]

May be the fluctuation of voltage on the cathode are used as to perform a fluctuation on G2 voltage in according with the cathode

That's the way I'm seeing it also, I think the cathode is out of phase with G2 so it's acting like NFB? 

There's a big filter cap on the cathode and the screens. No feedback going on here. It's actually fixed bias. Looks like this amp uses a combination of fixed and cathode bias.

[VMS]

It's actually fixed bias. Looks like this amp uses a combination of fixed and cathode bias.

No, it elevates the heater voltage.

[sluckey]

It's actually fixed bias. Looks like this amp uses a combination of fixed and cathode bias.

No, it elevates the heater voltage.

Heaters are elevated simply because they are connected to the cathodes. That connection to the screens makes a fixed voltage divider to the cathodes which will affect the bias of the output tubes.

[tubenit]

Franco,  I am thinking the 100k "mixing" resistor is needed because otherwise you will have a .01 & .002 in series giving you a .0016 value there.  I also think the 100k may add some smoothness to the overdrive following the 5879.

With respect, Tubenit

[Willabe]

That connection to the screens makes a fixed voltage divider to the cathodes which will affect the bias of the output tubes.

How stable is the fixed voltage divider bias?

What happens when the power tubes draw more and/or less current? Will the K bias part and the fvd part rise/fall?

Will they fight each other or will they null each other out and keep the output tubes bias more stable?

Or, will it act like a compressor or act like an expander?



                   Brad       

[VMS]

It's actually fixed bias. Looks like this amp uses a combination of fixed and cathode bias.

No, it elevates the heater voltage.

Heaters are elevated simply because they are connected to the cathodes. That connection to the screens makes a fixed voltage divider to the cathodes which will affect the bias of the output tubes.

Oh, I see. I thought you meant the traditional fixed-bias with the negative voltage on the grids. My bad, sorry.

So the 25k and 250 resistors form the voltage divider and this gives the cathodes about 3,6 volts, right?

[sluckey]

3.6V is correct if there is no current thru the tubes. But there is current thru the tubes so the additional current thru the cathode resistor causes a much larger voltage drop. That accounts for the 30V listed in the tube voltage chart. I think it's a silly thing to do. Probably the same engineers that did that zero volt grid cathodyne in the GA-5.   

[kagliostro]

First, I apologize, I forgot to thank PRR for its excellent explanation 

MANY THANKS PRR

If you changed to self-bias and stiffened-up the G2 supply, then you *could* go down to 4K load, get more power, and (when maxxed-out) a very different amp.

I would like to have the circuit working as the original, no interest on having more power, I've interest only in that tone and 25W are enough

---

About the real function of the connection of the 25k resistor with the 250R cathode resistor now I think it will be something like Steve say (Thanks), but honestly I haven't yet catch this thing 100% and I must think further to it to understand

Steve, I'm still writing this post and I read your last post

That accounts for the 30V listed in the tube voltage chart. I think it's a silly thing to do. Probably the same engineers that did that zero volt grid cathodyne in the GA-5. 

don't put on me further doubts ( --  )

I was thinking to this thing as part of the Mojo of this amp, isn't it 

---

@ Tubenit

Franco,  I am thinking the 100k "mixing" resistor is needed because otherwise you will have a .01 & .002 in series giving you a .0016 value there.  I also think the 100k may add some smoothness to the overdrive following the 5879.

Yes, you are right, will be worthwhile to study this thing from the practical point of view testing it, in fact what you say is more than sensible

may be the simplest solution is to add the 100k I missed or study the right value for a single cap followed (or preceded) by a resistor

it will be interesting also to go deeper on this thing from the theoretical point of view, an in series capacitor forms a frequency filter in junction with the other values of the circuit where it is installed, the presence of an in series resistor sure alter his effect

---

Many Thanks to ALL (very interesting tread !)

Franco

[shooter]

Will the K bias part and the fvd part rise/fall?

On a straight K biased doesn't the rise/fall voltage/current act kinda like nfb for G1? or is it just a plate cathode relationship?
Am I correct in "diagraming" the circuit as 2 || paths, the normal Kr, tube, B+, and a 2nd Kr, 25k, G2, plate
Thanks K for helping our brains grow

[2deaf]

May be the fluctuation of voltage on the cathode are used as to perform a fluctuation on G2 voltage

That's what I was thinking, also.  As the current increases in the tube, the voltage drop across the 4K resistor decreases making the screen voltage higher.  But at the same time, the current through the screen circuit increases and that circuit is in parallel to the 25K resistor so this causes an increased voltage drop across the 4K resistor.  If the two opposing voltage drops are similar, they will hold the screen voltage constant.

Even if this is correct, it's still a silly thing to do in a guitar amp.

[kagliostro]

Thanks K for helping our brains grow

I really must thank ALL of you for the given help

I really like when I learn something new and I'm happy when I discover something of interesting concealed in an old schematic

and Gibson schematics looks to be really interestings

I didn't catch this thing entirely till now, but I hope I'll arrive to understand it better

Franco

[shooter]

Gibson schematics looks to be really interesting

I think most purists see Gibson amps as the odd duck.  The last upgrade amp I did uses a tranny PI, taken from a skylark schematic.  I might try your 0 volt G1 PI next 

[PRR]

> Is the 25K R from the screens to the K's correct?

Well-spotted. And explains some bits I was confused on.

> I think the cathode is out of phase with G2 so it's acting like NFB?

It is a push-pull amp. "Both phases" at both points.

Older 6L6 had a low G2 rating. We often want a lower voltage on it. A dropping resistor does not work well, because 6L6 G2 current can vary a LOT, causing varying and uncertain voltage drop.

The 25K is a bleeder to swamp the G2 current and get more-consistent voltage drop in the 4K resistor.

Since G2 current comes out the cathode, the logical place to return the bottom of the 25K *is* to the cathode resistor.

> How stable is the fixed voltage divider bias?

Not very. Bleeder current is only 1/10th of the total. It avoids Vg2 soaring at idle; it has very little effect on with-signal operation. It's not a magic trick, just good (if wasteful) engineering.

You find the same idea on old Hammond organs, which also needed a semi-constant DC current for the field-coils of the several speakers.

To make this plan FIXED bias the bleeder current would have to be much-much-more than tube current. Much more heat in the resistors than in the tubes. PT about 3X larger than it "needs to be". Nobody does that.

[DummyLoad]

use these for 25W

http://www.tube-town.net/ttstore/Transformer/Outputtransformer/Marshall/Hammond-1750NA-Marshall-OT-JTM30::4635.html
http://www.tube-town.net/ttstore/Transformer/Power-Transformer/Audio/Hammond-373CZ::1011.html

--pete

[kagliostro]

Thanks PRR & Pete

The 25K is a bleeder to swamp the G2 current and get more-consistent voltage drop in the 4K resistor.

Since G2 current comes out the cathode, the logical place to return the bottom of the 25K *is* to the cathode resistor.

Bleeder current is only 1/10th of the total. It avoids Vg2 soaring at idle; it has very little effect on with-signal operation.

It's not a magic trick, just good (if wasteful) engineering.

Very interesting, thanks again

Franco