A few questions about a 1-channel AB763 build's voltages

[markmalin]

Hi all!
I just completed an AB763 build using Doug's One Channel AB763 schematic, but with the following changes/notes:
- no Vibrato circuit
- added Mid pot, and Rob's tone stack lift mod (SPST switch for mid pot's ground connection)
- I'm using 6L6GC power tubes
- I'm using a GZ34S rectifier
- My PT is a Mercury spare I had, 325-0-325
- I've wired the power supply filter caps like the DR where the first two are in parallel (not series)
- I've used 2 10K 3W resistors for my dropping resistors to get the preamp plate voltages into this range
- current plate voltages: (listed below in plate a/plate b order "VDC")
V1 (first gain stage) : 191/185
V2 (reverb driver)    : 409/409
V3 (rev recovery)     : 198/196
V4 (phase inverter)  : 199/196
V5/V6 power tubes  :  410/410
- bias notes:
-67.9 VDC seen at the bias pot input side
-47.8 VDC at the bias pot output (where the two 220k resistors are shown)
25.5 mA is the maximum I can dial up the bias to (measured using a 1 Ohm resistor between pin 8 and ground)
 
The amp works, sounds Ok, but I have a few questions/concerns I'm trying to understand.


1.) My PT is 325-0-325.  I'm seeing about 624V AC at the rectifier input pins (i.e. the feed from the transformer) and I can only get 410 VDC at the 6L6GC's plates.  Is this an OK operating range for 6L6GC's?  They sound Ok, though I can't get the bias current set any higher than 25 mA.  My hope is I can just use this transformer and this is sufficient plate voltage for the power tubes to sound good.  (FWIW I will be running this amp primarily at low volumes, live, for jazz)


2.) On an AB763, are my preamp plate voltages still too high?  (see above).  I see Fender schematics have them more in the 170'ish range


3.) Do I need to modify the 22k resistor on the bias pot (from one leg to ground) to allow me to be able to increase the bias current?


4.) Is the GZ34S rectifier the right one for this application, or should I use a different rectifier?


Humbly,
Mark

[markmalin]

I can post better pictures of the chassis if needed
Mark

[shooter]

-47.8 VDC at the bias pot output

what did you measure without PA tubes for "range", -27 to -60 kinda thing

[shooter]

the schematic I have shows range R

[tubeswell]

410V x 0.025A = 10W. 6L6GCs would sound better at 18-23W


Simplest experiment to begin with is decrease the amount of tail* resistance in the bias supply. Solder-tack another resistor in parallel, and see where you land.


*the resistor between the bias adjustment pot and the ground.


Re: the pre-amp plate voltages - a 20V difference is nothing to be concerned about.

[2deaf]

Change R67 (470) to 1K to get more mA's in idle current.

[markmalin]

the schematic I have shows range R

So shooter, you're wondering about voltage at either side of R67 with the power amp tubes removed?

[glass54]

Hi Mark
I built the AB763 1 Channel with Hoffman Turret Board. Used my own Toroidal Tx with 40AV Bias winding, BUT I used a slight variation of Bias circuit (attached). In my case I used 6V6 outputs and R36 at 22k to suit. I suggest that you use R36 at 10k as per attached. This circuit was used by Sluckey in his AB763 De Lite and is my preferred 2 stage filter for Bias Supply. Your Target voltage is lower obviously to suit 6L6 Bias requirements. And try it with no output tubes to get the resistor/pot in the correct range  .
Kind Regards
Mirek

[markmalin]

Thanks for the input, guys.  I modified the bias range -- in this case it was quicker to swap out the bias range resistor (I had a 1K 3W handy).  I'm able to set the 6L6GC's to about 32mA and the amp sounds great.


I won't sweat the higher plate voltages on the preamp tubes (thanks tubeswell).
So the 6L6's are running at only 10W...hm...  When I have a little extra money and time I may pick up another power transformer and change it out.  I hate to lose the Mercury.


Mark

[tubeswell]

Thanks for the input, guys.  I modified the bias range -- in this case it was quicker to swap out the bias range resistor (I had a 1K 3W handy).  I'm able to set the 6L6GC's to about 32mA and the amp sounds great.


I won't sweat the higher plate voltages on the preamp tubes (thanks tubeswell).
So the 6L6's are running at only 10W...hm...  When I have a little extra money and time I may pick up another power transformer and change it out.  I hate to lose the Mercury.


Mark

Well if you've got them up to 32mA, it's probably more like 410V x 0.032A* = 13W


*32mA

[markmalin]

Thanks for the input, guys.  I modified the bias range -- in this case it was quicker to swap out the bias range resistor (I had a 1K 3W handy).  I'm able to set the 6L6GC's to about 32mA and the amp sounds great.


I won't sweat the higher plate voltages on the preamp tubes (thanks tubeswell).
So the 6L6's are running at only 10W...hm...  When I have a little extra money and time I may pick up another power transformer and change it out.  I hate to lose the Mercury.


Mark

Well if you've got them up to 32mA, it's probably more like 410V x 0.032A* = 13W


*32mA

Thanks!  That's helpful. 

[PRR]

> PT is 325-0-325.  I'm seeing about 624V AC at the rectifier input pins (i.e. the feed from the transformer) and I can only get 410 VDC at the 6L6GC

Why "only"? 325V AC times 1.414 is 460V. Allow 50V drop in a vacuum rectifier is 410V, as you see.

410V is a bit high for original recipe non-GC 6L6 but entirely OK for any "6L6" made in the last 50 years.

25mA is cold and 35mA not a lot better. This does not affect maximum power but how smoothly or hoarsely it rises from idle to medium voice. You do NOT need to change the PT to fix this.

[2deaf]

So while leaving the 1K resistor at R67, try what tubeswell said and reduce R66 (22K).  Figuring in some sag with higher idle current, you might need a bias voltage around -37V to get 70% dissipation.  Replacing R66 (22K) with a 10K resistor could well give you a range from -29V to -59V.   

[glass54]

Mark
I would keep your Mercury Tx.
The original Deluxe Rev BF (6V6) had a nominal Node A supply of +420V. The AB763 Vibrolux (6L6) had nominal +415V at A node. AB763 Pro Amp (6L6) had +440V and +375V at node D (preamp supply). Mesa ran a lot of their input tubes close to +400V 
If your happy with the amp, the only thing I would do is possibly bridge R36 (22k) with another 22k or 33k to extend your bias range and get the bias current to 38ma or so. You HT will probably drop by a few volts.
Anyway, enjoy your amp 
Kind regards
Mirek

[markmalin]

Thanks, guys. I'll change out the 22k resistor on the bias pot to give me more range and try to get the bias more in line.


PRR - I was reading about vintage style 6L6 vs 6L6GC and the maximum plate voltages.  I'm using GC's, but that's something I had in mind, the possibility of using regular 6L6's.  In any case, I'll keep playing with the bias.


Mark

[tubeswell]

I'll change out the 22k resistor on the bias pot to give me more range and try to get the bias more in line.

Handy tip* - Until you find the 'happy place', its quicker to solder-tack another resistor in parallel. That way if you want to alter it further you can untack that one and tack other one in. (Hint: resistances in parallel = lower resistance

)


*repeating myself

[markmalin]

Handy tip* - Until you find the 'happy place', its quicker to solder-tack another resistor in parallel. That way if you want to alter it further you can untack that one and tack other one in. (Hint: resistances in parallel = lower resistance
*repeating myself

Thanks :)

[2deaf]

The 6L6GC is rated at 30W, so 70% would be 21W.  To find the necessary current, divide 21W by the plate voltage, e.g. 21W / 400V = 52.5mA. 

[markmalin]

The 6L6GC is rated at 30W, so 70% would be 21W.  To find the necessary current, divide 21W by the plate voltage, e.g. 21W / 400V = 52.5mA.

So the goal is around 70%, which I get.  People often talk about 35-38 mA for 6L6's as kind of a ballpark figure.  I realize that's one of those generalizations you read.  Using Robinette's bias calculator, in the 50 mA is what I'm aiming for to get within the 70% range, but is that considered a "lot of current" just idling?  Or maybe not all that much more than you'd see in a standard Class A amp?  Just thinking out loud here...


Mark

[2deaf]

So the goal is around 70%, which I get.  People often talk about 35-38 mA for 6L6's as kind of a ballpark figure.  I realize that's one of those generalizations you read.  Using Robinette's bias calculator, in the 50 mA is what I'm aiming for to get within the 70% range, but is that considered a "lot of current" just idling?  Or maybe not all that much more than you'd see in a standard Class A amp?

I don't consider 50mA to be a lot of current as long as the voltage isn't too high.  It's the wattage that matters and wattage is current times voltage.



 

[markmalin]

So the goal is around 70%, which I get.  People often talk about 35-38 mA for 6L6's as kind of a ballpark figure.  I realize that's one of those generalizations you read.  Using Robinette's bias calculator, in the 50 mA is what I'm aiming for to get within the 70% range, but is that considered a "lot of current" just idling?  Or maybe not all that much more than you'd see in a standard Class A amp?

I don't consider 50mA to be a lot of current as long as the voltage isn't too high.  It's the wattage that matters and wattage is current times voltage.

That makes total sense. 

[Willabe]

2.) On an AB763, are my preamp plate voltages still too high?  (see above).  I see Fender schematics have them more in the 170'ish range.

The AB763 Fender 6L6GC tube amps had higher dcv's on the preamp tubes, Bassmen, Concert, Pro, Twin, etc.

Super Reverb amps have 270dcv on the 1st 12AX7 plates.

[markmalin]

2.) On an AB763, are my preamp plate voltages still too high?  (see above).  I see Fender schematics have them more in the 170'ish range.

The AB763 Fender 6L6GC tube amps had higher dcv's on the preamp tubes, Bassmen, Concert, Pro, Twin, etc.

Super Reverb amps have 270dcv on the 1st 12AX7 plates.

Interesting...  I hadn't looked at the bigger amps, just the DR. 

[PRR]

> a "lot of current" just idling?

The peak LOUD current for 400V and 4Kct load is about 300mA in one tube. This will give some gain. For very-soft near-idle sounds we want a similar gain, but now both tubes are helping. An approximation of the gain/current curve suggests the idle per tube should be 1/4 the peak current. So 75mA.

400V*75mA is 30 Watts per tube. That's barely legal, IMHO not a good op-point, and leaves no room for medium-loud power output.

25mA idle is probably noticeably harsh on very soft sounds.

50mA is halfway from too-cold 25mA to too-hot 75mA. Depending on other factors, 35mA or 60mA may be perfectly fine too.

I once ran the bias on a hi-fi amplifier intended to bias cool. My ears had a small preference for about 150%-200% of rated Pdiss. That would be bad for the tubes and 50% Pdiss did not sound that much different. There's no magic spot where it "comes into focus". Teaching "70% rule" avoids long pointless discussions.

[markmalin]

> a "lot of current" just idling?

The peak LOUD current for 400V and 4Kct load is about 300mA in one tube. This will give some gain. For very-soft near-idle sounds we want a similar gain, but now both tubes are helping. An approximation of the gain/current curve suggests the idle per tube should be 1/4 the peak current. So 75mA.

400V*75mA is 30 Watts per tube. That's barely legal, IMHO not a good op-point, and leaves no room for medium-loud power output.

25mA idle is probably noticeably harsh on very soft sounds.

50mA is halfway from too-cold 25mA to too-hot 75mA. Depending on other factors, 35mA or 60mA may be perfectly fine too.

I once ran the bias on a hi-fi amplifier intended to bias cool. My ears had a small preference for about 150%-200% of rated Pdiss. That would be bad for the tubes and 50% Pdiss did not sound that much different. There's no magic spot where it "comes into focus". Teaching "70% rule" avoids long pointless discussions.

I know I've read several people saying to just use your ear.  But all this discussion has been really helpful.  I tend to be kind of a "literal" guy, so this is has been a learning process on this project.

[markmalin]

At the risk of taking this discussion in a different direction, I was playing around a bit with V1.


I had bought an EH 12AX7 and a new Mullard 12AX7 to try in V1.  I prefer the tone of the Mullard, but it's super microphonic.  Annoyingly so, so I put the EH back in.  Rob Robinette has a suggestion to try a 12AY7 (like in the 5E3) for the first gain stage in an AB763 to reduce gain.  I had a spare Mullard 12AY7, put it in there, and I really like the tone, though it's surprising how much less volume there is.  Also, the plate voltages on the 12AY7 dropped down considerably.  IIRC, something like 160.  Is that because there is a different resistance in the 12AY7 vs AX, or does it draw more current (giving a lower plate voltage)?  I like the behaviour of the amp now for jazz, just curious.


I guess my question is, should I change anything on V1 if I keep the AY in there?  Like change where it is biased?

[shooter]

fwiw;
I like n use AU7 in V1, plate voltage rarely gets to 70vdc but I'm amplifying a .2v signal so I don't need much "swing" range.

if you have the typical 1.5k 100k setup and want to play, start with 2.2k cathode 120k plate.
to "get back" some gain up the bypass cap by "the next value"  so If you have 2.2uf, try 4.7uf.  anything past 15uf and mud starts leaking in

[2deaf]

I guess my question is, should I change anything on V1 if I keep the AY in there?  Like change where it is biased?

I just read that Merlin article that shooter linked in some other thread and drew some load lines for a 12AY7 with a 100K plate resistor and a 1.5K cathode resistor.  The bias point is beautifully centre-biased with those values at an E_bb of 290V, so I wouldn't change anything on V1.  Looks like it's about half the gain of a 12AX7 with lots of headroom.

[shooter]

drew some load lines

once I did that, now I steal and scope for flavor 

[markmalin]

I guess my question is, should I change anything on V1 if I keep the AY in there?  Like change where it is biased?

I just read that Merlin article that shooter linked in some other thread and drew some load lines for a 12AY7 with a 100K plate resistor and a 1.5K cathode resistor.  The bias point is beautifully centre-biased with those values at an E_bb of 290V, so I wouldn't change anything on V1.  Looks like it's about half the gain of a 12AX7 with lots of headroom.

Thanks, 2deaf!!  I do like the way it's sounding :) 


Mark